├── .github └── FUNDING.yml ├── LICENSE ├── ML_model.py ├── README.md ├── bert_model_sen.py ├── calc_AUC.py ├── cnn_model_sen.py ├── construct.py ├── construct_test.py ├── construct_train.py ├── data ├── right.csv ├── submit_example.csv ├── test.feature.csv └── train.news.csv ├── lstm_model.py ├── requirements.txt └── web.py /.github/FUNDING.yml: -------------------------------------------------------------------------------- 1 | # These are supported funding model platforms 2 | 3 | github: # Replace with up to 4 GitHub Sponsors-enabled usernames e.g., [user1, user2] 4 | patreon: # Replace with a single Patreon username 5 | open_collective: # Replace with a single Open Collective username 6 | ko_fi: # Replace with a single Ko-fi username 7 | tidelift: # Replace with a single Tidelift platform-name/package-name e.g., npm/babel 8 | community_bridge: # Replace with a single Community Bridge project-name e.g., cloud-foundry 9 | liberapay: # Replace with a single Liberapay username 10 | issuehunt: # Replace with a single IssueHunt username 11 | lfx_crowdfunding: # Replace with a single LFX Crowdfunding project-name e.g., cloud-foundry 12 | polar: # Replace with a single Polar username 13 | buy_me_a_coffee: # Replace with a single Buy Me a Coffee username 14 | thanks_dev: # Replace with a single thanks.dev username 15 | custom: ['https://pay.lazyforever.top/'] 16 | # Replace with up to 4 custom sponsorship URLs e.g., ['link1', 'link2'] 17 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | GNU AFFERO GENERAL PUBLIC LICENSE 2 | Version 3, 19 November 2007 3 | 4 | Copyright (C) 2007 Free Software Foundation, Inc. 5 | Everyone is permitted to copy and distribute verbatim copies 6 | of this license document, but changing it is not allowed. 7 | 8 | Preamble 9 | 10 | The GNU Affero General Public License is a free, copyleft license for 11 | software and other kinds of works, specifically designed to ensure 12 | cooperation with the community in the case of network server software. 13 | 14 | The licenses for most software and other practical works are designed 15 | to take away your freedom to share and change the works. By contrast, 16 | our General Public Licenses are intended to guarantee your freedom to 17 | share and change all versions of a program--to make sure it remains free 18 | software for all its users. 19 | 20 | When we speak of free software, we are referring to freedom, not 21 | price. Our General Public Licenses are designed to make sure that you 22 | have the freedom to distribute copies of free software (and charge for 23 | them if you wish), that you receive source code or can get it if you 24 | want it, that you can change the software or use pieces of it in new 25 | free programs, and that you know you can do these things. 26 | 27 | Developers that use our General Public Licenses protect your rights 28 | with two steps: (1) assert copyright on the software, and (2) offer 29 | you this License which gives you legal permission to copy, distribute 30 | and/or modify the software. 31 | 32 | A secondary benefit of defending all users' freedom is that 33 | improvements made in alternate versions of the program, if they 34 | receive widespread use, become available for other developers to 35 | incorporate. Many developers of free software are heartened and 36 | encouraged by the resulting cooperation. However, in the case of 37 | software used on network servers, this result may fail to come about. 38 | The GNU General Public License permits making a modified version and 39 | letting the public access it on a server without ever releasing its 40 | source code to the public. 41 | 42 | The GNU Affero General Public License is designed specifically to 43 | ensure that, in such cases, the modified source code becomes available 44 | to the community. It requires the operator of a network server to 45 | provide the source code of the modified version running there to the 46 | users of that server. Therefore, public use of a modified version, on 47 | a publicly accessible server, gives the public access to the source 48 | code of the modified version. 49 | 50 | An older license, called the Affero General Public License and 51 | published by Affero, was designed to accomplish similar goals. This is 52 | a different license, not a version of the Affero GPL, but Affero has 53 | released a new version of the Affero GPL which permits relicensing under 54 | this license. 55 | 56 | The precise terms and conditions for copying, distribution and 57 | modification follow. 58 | 59 | TERMS AND CONDITIONS 60 | 61 | 0. Definitions. 62 | 63 | "This License" refers to version 3 of the GNU Affero General Public License. 64 | 65 | "Copyright" also means copyright-like laws that apply to other kinds of 66 | works, such as semiconductor masks. 67 | 68 | "The Program" refers to any copyrightable work licensed under this 69 | License. Each licensee is addressed as "you". "Licensees" and 70 | "recipients" may be individuals or organizations. 71 | 72 | To "modify" a work means to copy from or adapt all or part of the work 73 | in a fashion requiring copyright permission, other than the making of an 74 | exact copy. The resulting work is called a "modified version" of the 75 | earlier work or a work "based on" the earlier work. 76 | 77 | A "covered work" means either the unmodified Program or a work based 78 | on the Program. 79 | 80 | To "propagate" a work means to do anything with it that, without 81 | permission, would make you directly or secondarily liable for 82 | infringement under applicable copyright law, except executing it on a 83 | computer or modifying a private copy. Propagation includes copying, 84 | distribution (with or without modification), making available to the 85 | public, and in some countries other activities as well. 86 | 87 | To "convey" a work means any kind of propagation that enables other 88 | parties to make or receive copies. Mere interaction with a user through 89 | a computer network, with no transfer of a copy, is not conveying. 90 | 91 | An interactive user interface displays "Appropriate Legal Notices" 92 | to the extent that it includes a convenient and prominently visible 93 | feature that (1) displays an appropriate copyright notice, and (2) 94 | tells the user that there is no warranty for the work (except to the 95 | extent that warranties are provided), that licensees may convey the 96 | work under this License, and how to view a copy of this License. If 97 | the interface presents a list of user commands or options, such as a 98 | menu, a prominent item in the list meets this criterion. 99 | 100 | 1. Source Code. 101 | 102 | The "source code" for a work means the preferred form of the work 103 | for making modifications to it. "Object code" means any non-source 104 | form of a work. 105 | 106 | A "Standard Interface" means an interface that either is an official 107 | standard defined by a recognized standards body, or, in the case of 108 | interfaces specified for a particular programming language, one that 109 | is widely used among developers working in that language. 110 | 111 | The "System Libraries" of an executable work include anything, other 112 | than the work as a whole, that (a) is included in the normal form of 113 | packaging a Major Component, but which is not part of that Major 114 | Component, and (b) serves only to enable use of the work with that 115 | Major Component, or to implement a Standard Interface for which an 116 | implementation is available to the public in source code form. A 117 | "Major Component", in this context, means a major essential component 118 | (kernel, window system, and so on) of the specific operating system 119 | (if any) on which the executable work runs, or a compiler used to 120 | produce the work, or an object code interpreter used to run it. 121 | 122 | The "Corresponding Source" for a work in object code form means all 123 | the source code needed to generate, install, and (for an executable 124 | work) run the object code and to modify the work, including scripts to 125 | control those activities. However, it does not include the work's 126 | System Libraries, or general-purpose tools or generally available free 127 | programs which are used unmodified in performing those activities but 128 | which are not part of the work. For example, Corresponding Source 129 | includes interface definition files associated with source files for 130 | the work, and the source code for shared libraries and dynamically 131 | linked subprograms that the work is specifically designed to require, 132 | such as by intimate data communication or control flow between those 133 | subprograms and other parts of the work. 134 | 135 | The Corresponding Source need not include anything that users 136 | can regenerate automatically from other parts of the Corresponding 137 | Source. 138 | 139 | The Corresponding Source for a work in source code form is that 140 | same work. 141 | 142 | 2. Basic Permissions. 143 | 144 | All rights granted under this License are granted for the term of 145 | copyright on the Program, and are irrevocable provided the stated 146 | conditions are met. This License explicitly affirms your unlimited 147 | permission to run the unmodified Program. The output from running a 148 | covered work is covered by this License only if the output, given its 149 | content, constitutes a covered work. This License acknowledges your 150 | rights of fair use or other equivalent, as provided by copyright law. 151 | 152 | You may make, run and propagate covered works that you do not 153 | convey, without conditions so long as your license otherwise remains 154 | in force. You may convey covered works to others for the sole purpose 155 | of having them make modifications exclusively for you, or provide you 156 | with facilities for running those works, provided that you comply with 157 | the terms of this License in conveying all material for which you do 158 | not control copyright. Those thus making or running the covered works 159 | for you must do so exclusively on your behalf, under your direction 160 | and control, on terms that prohibit them from making any copies of 161 | your copyrighted material outside their relationship with you. 162 | 163 | Conveying under any other circumstances is permitted solely under 164 | the conditions stated below. Sublicensing is not allowed; section 10 165 | makes it unnecessary. 166 | 167 | 3. Protecting Users' Legal Rights From Anti-Circumvention Law. 168 | 169 | No covered work shall be deemed part of an effective technological 170 | measure under any applicable law fulfilling obligations under article 171 | 11 of the WIPO copyright treaty adopted on 20 December 1996, or 172 | similar laws prohibiting or restricting circumvention of such 173 | measures. 174 | 175 | When you convey a covered work, you waive any legal power to forbid 176 | circumvention of technological measures to the extent such circumvention 177 | is effected by exercising rights under this License with respect to 178 | the covered work, and you disclaim any intention to limit operation or 179 | modification of the work as a means of enforcing, against the work's 180 | users, your or third parties' legal rights to forbid circumvention of 181 | technological measures. 182 | 183 | 4. Conveying Verbatim Copies. 184 | 185 | You may convey verbatim copies of the Program's source code as you 186 | receive it, in any medium, provided that you conspicuously and 187 | appropriately publish on each copy an appropriate copyright notice; 188 | keep intact all notices stating that this License and any 189 | non-permissive terms added in accord with section 7 apply to the code; 190 | keep intact all notices of the absence of any warranty; and give all 191 | recipients a copy of this License along with the Program. 192 | 193 | You may charge any price or no price for each copy that you convey, 194 | and you may offer support or warranty protection for a fee. 195 | 196 | 5. Conveying Modified Source Versions. 197 | 198 | You may convey a work based on the Program, or the modifications to 199 | produce it from the Program, in the form of source code under the 200 | terms of section 4, provided that you also meet all of these conditions: 201 | 202 | a) The work must carry prominent notices stating that you modified 203 | it, and giving a relevant date. 204 | 205 | b) The work must carry prominent notices stating that it is 206 | released under this License and any conditions added under section 207 | 7. This requirement modifies the requirement in section 4 to 208 | "keep intact all notices". 209 | 210 | c) You must license the entire work, as a whole, under this 211 | License to anyone who comes into possession of a copy. This 212 | License will therefore apply, along with any applicable section 7 213 | additional terms, to the whole of the work, and all its parts, 214 | regardless of how they are packaged. This License gives no 215 | permission to license the work in any other way, but it does not 216 | invalidate such permission if you have separately received it. 217 | 218 | d) If the work has interactive user interfaces, each must display 219 | Appropriate Legal Notices; however, if the Program has interactive 220 | interfaces that do not display Appropriate Legal Notices, your 221 | work need not make them do so. 222 | 223 | A compilation of a covered work with other separate and independent 224 | works, which are not by their nature extensions of the covered work, 225 | and which are not combined with it such as to form a larger program, 226 | in or on a volume of a storage or distribution medium, is called an 227 | "aggregate" if the compilation and its resulting copyright are not 228 | used to limit the access or legal rights of the compilation's users 229 | beyond what the individual works permit. Inclusion of a covered work 230 | in an aggregate does not cause this License to apply to the other 231 | parts of the aggregate. 232 | 233 | 6. Conveying Non-Source Forms. 234 | 235 | You may convey a covered work in object code form under the terms 236 | of sections 4 and 5, provided that you also convey the 237 | machine-readable Corresponding Source under the terms of this License, 238 | in one of these ways: 239 | 240 | a) Convey the object code in, or embodied in, a physical product 241 | (including a physical distribution medium), accompanied by the 242 | Corresponding Source fixed on a durable physical medium 243 | customarily used for software interchange. 244 | 245 | b) Convey the object code in, or embodied in, a physical product 246 | (including a physical distribution medium), accompanied by a 247 | written offer, valid for at least three years and valid for as 248 | long as you offer spare parts or customer support for that product 249 | model, to give anyone who possesses the object code either (1) a 250 | copy of the Corresponding Source for all the software in the 251 | product that is covered by this License, on a durable physical 252 | medium customarily used for software interchange, for a price no 253 | more than your reasonable cost of physically performing this 254 | conveying of source, or (2) access to copy the 255 | Corresponding Source from a network server at no charge. 256 | 257 | c) Convey individual copies of the object code with a copy of the 258 | written offer to provide the Corresponding Source. This 259 | alternative is allowed only occasionally and noncommercially, and 260 | only if you received the object code with such an offer, in accord 261 | with subsection 6b. 262 | 263 | d) Convey the object code by offering access from a designated 264 | place (gratis or for a charge), and offer equivalent access to the 265 | Corresponding Source in the same way through the same place at no 266 | further charge. You need not require recipients to copy the 267 | Corresponding Source along with the object code. If the place to 268 | copy the object code is a network server, the Corresponding Source 269 | may be on a different server (operated by you or a third party) 270 | that supports equivalent copying facilities, provided you maintain 271 | clear directions next to the object code saying where to find the 272 | Corresponding Source. Regardless of what server hosts the 273 | Corresponding Source, you remain obligated to ensure that it is 274 | available for as long as needed to satisfy these requirements. 275 | 276 | e) Convey the object code using peer-to-peer transmission, provided 277 | you inform other peers where the object code and Corresponding 278 | Source of the work are being offered to the general public at no 279 | charge under subsection 6d. 280 | 281 | A separable portion of the object code, whose source code is excluded 282 | from the Corresponding Source as a System Library, need not be 283 | included in conveying the object code work. 284 | 285 | A "User Product" is either (1) a "consumer product", which means any 286 | tangible personal property which is normally used for personal, family, 287 | or household purposes, or (2) anything designed or sold for incorporation 288 | into a dwelling. In determining whether a product is a consumer product, 289 | doubtful cases shall be resolved in favor of coverage. For a particular 290 | product received by a particular user, "normally used" refers to a 291 | typical or common use of that class of product, regardless of the status 292 | of the particular user or of the way in which the particular user 293 | actually uses, or expects or is expected to use, the product. A product 294 | is a consumer product regardless of whether the product has substantial 295 | commercial, industrial or non-consumer uses, unless such uses represent 296 | the only significant mode of use of the product. 297 | 298 | "Installation Information" for a User Product means any methods, 299 | procedures, authorization keys, or other information required to install 300 | and execute modified versions of a covered work in that User Product from 301 | a modified version of its Corresponding Source. The information must 302 | suffice to ensure that the continued functioning of the modified object 303 | code is in no case prevented or interfered with solely because 304 | modification has been made. 305 | 306 | If you convey an object code work under this section in, or with, or 307 | specifically for use in, a User Product, and the conveying occurs as 308 | part of a transaction in which the right of possession and use of the 309 | User Product is transferred to the recipient in perpetuity or for a 310 | fixed term (regardless of how the transaction is characterized), the 311 | Corresponding Source conveyed under this section must be accompanied 312 | by the Installation Information. But this requirement does not apply 313 | if neither you nor any third party retains the ability to install 314 | modified object code on the User Product (for example, the work has 315 | been installed in ROM). 316 | 317 | The requirement to provide Installation Information does not include a 318 | requirement to continue to provide support service, warranty, or updates 319 | for a work that has been modified or installed by the recipient, or for 320 | the User Product in which it has been modified or installed. Access to a 321 | network may be denied when the modification itself materially and 322 | adversely affects the operation of the network or violates the rules and 323 | protocols for communication across the network. 324 | 325 | Corresponding Source conveyed, and Installation Information provided, 326 | in accord with this section must be in a format that is publicly 327 | documented (and with an implementation available to the public in 328 | source code form), and must require no special password or key for 329 | unpacking, reading or copying. 330 | 331 | 7. Additional Terms. 332 | 333 | "Additional permissions" are terms that supplement the terms of this 334 | License by making exceptions from one or more of its conditions. 335 | Additional permissions that are applicable to the entire Program shall 336 | be treated as though they were included in this License, to the extent 337 | that they are valid under applicable law. If additional permissions 338 | apply only to part of the Program, that part may be used separately 339 | under those permissions, but the entire Program remains governed by 340 | this License without regard to the additional permissions. 341 | 342 | When you convey a copy of a covered work, you may at your option 343 | remove any additional permissions from that copy, or from any part of 344 | it. (Additional permissions may be written to require their own 345 | removal in certain cases when you modify the work.) You may place 346 | additional permissions on material, added by you to a covered work, 347 | for which you have or can give appropriate copyright permission. 348 | 349 | Notwithstanding any other provision of this License, for material you 350 | add to a covered work, you may (if authorized by the copyright holders of 351 | that material) supplement the terms of this License with terms: 352 | 353 | a) Disclaiming warranty or limiting liability differently from the 354 | terms of sections 15 and 16 of this License; or 355 | 356 | b) Requiring preservation of specified reasonable legal notices or 357 | author attributions in that material or in the Appropriate Legal 358 | Notices displayed by works containing it; or 359 | 360 | c) Prohibiting misrepresentation of the origin of that material, or 361 | requiring that modified versions of such material be marked in 362 | reasonable ways as different from the original version; or 363 | 364 | d) Limiting the use for publicity purposes of names of licensors or 365 | authors of the material; or 366 | 367 | e) Declining to grant rights under trademark law for use of some 368 | trade names, trademarks, or service marks; or 369 | 370 | f) Requiring indemnification of licensors and authors of that 371 | material by anyone who conveys the material (or modified versions of 372 | it) with contractual assumptions of liability to the recipient, for 373 | any liability that these contractual assumptions directly impose on 374 | those licensors and authors. 375 | 376 | All other non-permissive additional terms are considered "further 377 | restrictions" within the meaning of section 10. If the Program as you 378 | received it, or any part of it, contains a notice stating that it is 379 | governed by this License along with a term that is a further 380 | restriction, you may remove that term. If a license document contains 381 | a further restriction but permits relicensing or conveying under this 382 | License, you may add to a covered work material governed by the terms 383 | of that license document, provided that the further restriction does 384 | not survive such relicensing or conveying. 385 | 386 | If you add terms to a covered work in accord with this section, you 387 | must place, in the relevant source files, a statement of the 388 | additional terms that apply to those files, or a notice indicating 389 | where to find the applicable terms. 390 | 391 | Additional terms, permissive or non-permissive, may be stated in the 392 | form of a separately written license, or stated as exceptions; 393 | the above requirements apply either way. 394 | 395 | 8. Termination. 396 | 397 | You may not propagate or modify a covered work except as expressly 398 | provided under this License. Any attempt otherwise to propagate or 399 | modify it is void, and will automatically terminate your rights under 400 | this License (including any patent licenses granted under the third 401 | paragraph of section 11). 402 | 403 | However, if you cease all violation of this License, then your 404 | license from a particular copyright holder is reinstated (a) 405 | provisionally, unless and until the copyright holder explicitly and 406 | finally terminates your license, and (b) permanently, if the copyright 407 | holder fails to notify you of the violation by some reasonable means 408 | prior to 60 days after the cessation. 409 | 410 | Moreover, your license from a particular copyright holder is 411 | reinstated permanently if the copyright holder notifies you of the 412 | violation by some reasonable means, this is the first time you have 413 | received notice of violation of this License (for any work) from that 414 | copyright holder, and you cure the violation prior to 30 days after 415 | your receipt of the notice. 416 | 417 | Termination of your rights under this section does not terminate the 418 | licenses of parties who have received copies or rights from you under 419 | this License. If your rights have been terminated and not permanently 420 | reinstated, you do not qualify to receive new licenses for the same 421 | material under section 10. 422 | 423 | 9. Acceptance Not Required for Having Copies. 424 | 425 | You are not required to accept this License in order to receive or 426 | run a copy of the Program. Ancillary propagation of a covered work 427 | occurring solely as a consequence of using peer-to-peer transmission 428 | to receive a copy likewise does not require acceptance. However, 429 | nothing other than this License grants you permission to propagate or 430 | modify any covered work. These actions infringe copyright if you do 431 | not accept this License. Therefore, by modifying or propagating a 432 | covered work, you indicate your acceptance of this License to do so. 433 | 434 | 10. Automatic Licensing of Downstream Recipients. 435 | 436 | Each time you convey a covered work, the recipient automatically 437 | receives a license from the original licensors, to run, modify and 438 | propagate that work, subject to this License. You are not responsible 439 | for enforcing compliance by third parties with this License. 440 | 441 | An "entity transaction" is a transaction transferring control of an 442 | organization, or substantially all assets of one, or subdividing an 443 | organization, or merging organizations. If propagation of a covered 444 | work results from an entity transaction, each party to that 445 | transaction who receives a copy of the work also receives whatever 446 | licenses to the work the party's predecessor in interest had or could 447 | give under the previous paragraph, plus a right to possession of the 448 | Corresponding Source of the work from the predecessor in interest, if 449 | the predecessor has it or can get it with reasonable efforts. 450 | 451 | You may not impose any further restrictions on the exercise of the 452 | rights granted or affirmed under this License. For example, you may 453 | not impose a license fee, royalty, or other charge for exercise of 454 | rights granted under this License, and you may not initiate litigation 455 | (including a cross-claim or counterclaim in a lawsuit) alleging that 456 | any patent claim is infringed by making, using, selling, offering for 457 | sale, or importing the Program or any portion of it. 458 | 459 | 11. Patents. 460 | 461 | A "contributor" is a copyright holder who authorizes use under this 462 | License of the Program or a work on which the Program is based. The 463 | work thus licensed is called the contributor's "contributor version". 464 | 465 | A contributor's "essential patent claims" are all patent claims 466 | owned or controlled by the contributor, whether already acquired or 467 | hereafter acquired, that would be infringed by some manner, permitted 468 | by this License, of making, using, or selling its contributor version, 469 | but do not include claims that would be infringed only as a 470 | consequence of further modification of the contributor version. For 471 | purposes of this definition, "control" includes the right to grant 472 | patent sublicenses in a manner consistent with the requirements of 473 | this License. 474 | 475 | Each contributor grants you a non-exclusive, worldwide, royalty-free 476 | patent license under the contributor's essential patent claims, to 477 | make, use, sell, offer for sale, import and otherwise run, modify and 478 | propagate the contents of its contributor version. 479 | 480 | In the following three paragraphs, a "patent license" is any express 481 | agreement or commitment, however denominated, not to enforce a patent 482 | (such as an express permission to practice a patent or covenant not to 483 | sue for patent infringement). To "grant" such a patent license to a 484 | party means to make such an agreement or commitment not to enforce a 485 | patent against the party. 486 | 487 | If you convey a covered work, knowingly relying on a patent license, 488 | and the Corresponding Source of the work is not available for anyone 489 | to copy, free of charge and under the terms of this License, through a 490 | publicly available network server or other readily accessible means, 491 | then you must either (1) cause the Corresponding Source to be so 492 | available, or (2) arrange to deprive yourself of the benefit of the 493 | patent license for this particular work, or (3) arrange, in a manner 494 | consistent with the requirements of this License, to extend the patent 495 | license to downstream recipients. "Knowingly relying" means you have 496 | actual knowledge that, but for the patent license, your conveying the 497 | covered work in a country, or your recipient's use of the covered work 498 | in a country, would infringe one or more identifiable patents in that 499 | country that you have reason to believe are valid. 500 | 501 | If, pursuant to or in connection with a single transaction or 502 | arrangement, you convey, or propagate by procuring conveyance of, a 503 | covered work, and grant a patent license to some of the parties 504 | receiving the covered work authorizing them to use, propagate, modify 505 | or convey a specific copy of the covered work, then the patent license 506 | you grant is automatically extended to all recipients of the covered 507 | work and works based on it. 508 | 509 | A patent license is "discriminatory" if it does not include within 510 | the scope of its coverage, prohibits the exercise of, or is 511 | conditioned on the non-exercise of one or more of the rights that are 512 | specifically granted under this License. You may not convey a covered 513 | work if you are a party to an arrangement with a third party that is 514 | in the business of distributing software, under which you make payment 515 | to the third party based on the extent of your activity of conveying 516 | the work, and under which the third party grants, to any of the 517 | parties who would receive the covered work from you, a discriminatory 518 | patent license (a) in connection with copies of the covered work 519 | conveyed by you (or copies made from those copies), or (b) primarily 520 | for and in connection with specific products or compilations that 521 | contain the covered work, unless you entered into that arrangement, 522 | or that patent license was granted, prior to 28 March 2007. 523 | 524 | Nothing in this License shall be construed as excluding or limiting 525 | any implied license or other defenses to infringement that may 526 | otherwise be available to you under applicable patent law. 527 | 528 | 12. No Surrender of Others' Freedom. 529 | 530 | If conditions are imposed on you (whether by court order, agreement or 531 | otherwise) that contradict the conditions of this License, they do not 532 | excuse you from the conditions of this License. If you cannot convey a 533 | covered work so as to satisfy simultaneously your obligations under this 534 | License and any other pertinent obligations, then as a consequence you may 535 | not convey it at all. For example, if you agree to terms that obligate you 536 | to collect a royalty for further conveying from those to whom you convey 537 | the Program, the only way you could satisfy both those terms and this 538 | License would be to refrain entirely from conveying the Program. 539 | 540 | 13. Remote Network Interaction; Use with the GNU General Public License. 541 | 542 | Notwithstanding any other provision of this License, if you modify the 543 | Program, your modified version must prominently offer all users 544 | interacting with it remotely through a computer network (if your version 545 | supports such interaction) an opportunity to receive the Corresponding 546 | Source of your version by providing access to the Corresponding Source 547 | from a network server at no charge, through some standard or customary 548 | means of facilitating copying of software. This Corresponding Source 549 | shall include the Corresponding Source for any work covered by version 3 550 | of the GNU General Public License that is incorporated pursuant to the 551 | following paragraph. 552 | 553 | Notwithstanding any other provision of this License, you have 554 | permission to link or combine any covered work with a work licensed 555 | under version 3 of the GNU General Public License into a single 556 | combined work, and to convey the resulting work. The terms of this 557 | License will continue to apply to the part which is the covered work, 558 | but the work with which it is combined will remain governed by version 559 | 3 of the GNU General Public License. 560 | 561 | 14. Revised Versions of this License. 562 | 563 | The Free Software Foundation may publish revised and/or new versions of 564 | the GNU Affero General Public License from time to time. Such new versions 565 | will be similar in spirit to the present version, but may differ in detail to 566 | address new problems or concerns. 567 | 568 | Each version is given a distinguishing version number. If the 569 | Program specifies that a certain numbered version of the GNU Affero General 570 | Public License "or any later version" applies to it, you have the 571 | option of following the terms and conditions either of that numbered 572 | version or of any later version published by the Free Software 573 | Foundation. If the Program does not specify a version number of the 574 | GNU Affero General Public License, you may choose any version ever published 575 | by the Free Software Foundation. 576 | 577 | If the Program specifies that a proxy can decide which future 578 | versions of the GNU Affero General Public License can be used, that proxy's 579 | public statement of acceptance of a version permanently authorizes you 580 | to choose that version for the Program. 581 | 582 | Later license versions may give you additional or different 583 | permissions. However, no additional obligations are imposed on any 584 | author or copyright holder as a result of your choosing to follow a 585 | later version. 586 | 587 | 15. Disclaimer of Warranty. 588 | 589 | THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY 590 | APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT 591 | HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY 592 | OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, 593 | THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 594 | PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM 595 | IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF 596 | ALL NECESSARY SERVICING, REPAIR OR CORRECTION. 597 | 598 | 16. Limitation of Liability. 599 | 600 | IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING 601 | WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS 602 | THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY 603 | GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE 604 | USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF 605 | DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD 606 | PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), 607 | EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF 608 | SUCH DAMAGES. 609 | 610 | 17. Interpretation of Sections 15 and 16. 611 | 612 | If the disclaimer of warranty and limitation of liability provided 613 | above cannot be given local legal effect according to their terms, 614 | reviewing courts shall apply local law that most closely approximates 615 | an absolute waiver of all civil liability in connection with the 616 | Program, unless a warranty or assumption of liability accompanies a 617 | copy of the Program in return for a fee. 618 | 619 | END OF TERMS AND CONDITIONS 620 | 621 | How to Apply These Terms to Your New Programs 622 | 623 | If you develop a new program, and you want it to be of the greatest 624 | possible use to the public, the best way to achieve this is to make it 625 | free software which everyone can redistribute and change under these terms. 626 | 627 | To do so, attach the following notices to the program. It is safest 628 | to attach them to the start of each source file to most effectively 629 | state the exclusion of warranty; and each file should have at least 630 | the "copyright" line and a pointer to where the full notice is found. 631 | 632 | 633 | Copyright (C) 634 | 635 | This program is free software: you can redistribute it and/or modify 636 | it under the terms of the GNU Affero General Public License as published by 637 | the Free Software Foundation, either version 3 of the License, or 638 | (at your option) any later version. 639 | 640 | This program is distributed in the hope that it will be useful, 641 | but WITHOUT ANY WARRANTY; without even the implied warranty of 642 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 643 | GNU Affero General Public License for more details. 644 | 645 | You should have received a copy of the GNU Affero General Public License 646 | along with this program. If not, see . 647 | 648 | Also add information on how to contact you by electronic and paper mail. 649 | 650 | If your software can interact with users remotely through a computer 651 | network, you should also make sure that it provides a way for users to 652 | get its source. For example, if your program is a web application, its 653 | interface could display a "Source" link that leads users to an archive 654 | of the code. There are many ways you could offer source, and different 655 | solutions will be better for different programs; see section 13 for the 656 | specific requirements. 657 | 658 | You should also get your employer (if you work as a programmer) or school, 659 | if any, to sign a "copyright disclaimer" for the program, if necessary. 660 | For more information on this, and how to apply and follow the GNU AGPL, see 661 | . 662 | -------------------------------------------------------------------------------- /ML_model.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | import pandas as pd 3 | from threading import Thread 4 | from sklearn.model_selection import train_test_split 5 | from sklearn.feature_extraction.text import TfidfVectorizer 6 | from sklearn.svm import LinearSVC, SVC 7 | from sklearn.naive_bayes import MultinomialNB 8 | from sklearn.neighbors import KNeighborsClassifier 9 | from sklearn.tree import DecisionTreeClassifier 10 | from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier, AdaBoostClassifier 11 | from sklearn.neural_network import MLPClassifier 12 | from sklearn.linear_model import LogisticRegression 13 | import warnings 14 | warnings.filterwarnings("ignore") 15 | 16 | data=pd.read_csv("./data/train.news.csv") 17 | data_validate=pd.read_csv("./data/test.feature.csv") 18 | x,y,x_validate=data['Title'],data['label'],data_validate['Title'] 19 | print(x.shape,y.shape,x_validate.shape) 20 | print("读取数据完成") 21 | 22 | vectorizer=TfidfVectorizer() 23 | x=vectorizer.fit_transform(x) 24 | x_validate=vectorizer.transform(x_validate) 25 | x_train,x_test,y_train,y_test=train_test_split(x,y,test_size=0.2,stratify=y,random_state = 0) 26 | x_validate1,x_validate2=train_test_split(x_validate,test_size=0.5,random_state=0) 27 | 28 | # 线性支持向量机 29 | model1 = LinearSVC() 30 | # 朴素贝叶斯 31 | model2 = MultinomialNB() 32 | # K近邻 33 | model3 = KNeighborsClassifier(n_neighbors=50) 34 | # 决策树 35 | model4 = DecisionTreeClassifier(random_state=77) 36 | # 随机森林 37 | model5 = RandomForestClassifier(n_estimators=500, max_features='sqrt', random_state=10) 38 | # 梯度提升 39 | model6 = GradientBoostingClassifier(random_state=123) 40 | # 支持向量机 41 | model7 = SVC(kernel="rbf", random_state=77) 42 | # 神经网络 43 | model8 = MLPClassifier(hidden_layer_sizes=(16, 8), random_state=77, max_iter=10000) 44 | # AdaBoostClassifier 45 | model9 = AdaBoostClassifier(n_estimators=100, random_state=77) 46 | # 逻辑回归 47 | model10 = LogisticRegression(random_state=77) 48 | 49 | model_list = [model1, model2, model3, model4, model5, model6, model7, model8, model9, model10] 50 | model_name = ['线性支持向量机', '朴素贝叶斯', 'K近邻', '决策树', '随机森林', '梯度提升', '支持向量机', '神经网络', 'AdaBoostClassifier', '逻辑回归'] 51 | threads = [] 52 | 53 | def fit(model, name:str): 54 | model.fit(x_train, y_train) 55 | s=model.score(x_test, y_test) 56 | pre=model.predict(x_validate) 57 | np.savetxt(f"./data/机器学习{name}方法prediction.csv",pre,delimiter=',',fmt='%f') 58 | print(f'{name}方法在测试集的准确率为{s}') 59 | 60 | for i in range(len(model_list)): 61 | model=model_list[i] 62 | name=model_name[i] 63 | threads.append(Thread(target=fit, args=(model, name))) 64 | 65 | for thread in threads: 66 | thread.start() 67 | for thread in threads: 68 | thread.join() 69 | print("done!") -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Fake-news-detection 2 | 3 | 虚假新闻检测-南开大学《python语言程序设计》大作业，数据集来源[yaqingwang/WeFEND-AAAI20: Dataset for paper "Weak Supervision for Fake News Detection via Reinforcement Learning" published in AAAI'2020. (github.com)](https://github.com/yaqingwang/WeFEND-AAAI20) 4 | 5 | 本项目遵守[AGPL-3.0开源协议](LICENSE)，你可以自由地使用、修改、传播源代码，但是你必须遵守[AGPL-3.0许可证](LICENSE)的规定，即如果使用或修改我的源代码，需要注明作者并给出项目链接，并且在公共平台开源（例如在GitHub或GitLab等），否则我有权力进行追责。 6 | 7 | ***如果您觉得本项目还不错，请给出您的 star。另外，我以后可能还会将其他课程或项目的相关代码开源在 Github，欢迎 follow me。*** 8 | 9 | [![Star History Chart](https://api.star-history.com/svg?repos=lazy-forever/Fake-news-detection&type=Timeline)](https://star-history.com/#lazy-forever/Fake-news-detection&Timeline) 10 | 11 | ## 构建 12 | 13 | 确保你的电脑上已经装了pytorch 14 | 15 | ```shell 16 | pip install -r requirements.txt 17 | mkdir model 18 | ``` 19 | 20 | 下载[ymcui/Chinese-BERT-wwm: Pre-Training with Whole Word Masking for Chinese BERT（中文BERT-wwm系列模型）](https://github.com/ymcui/Chinese-BERT-wwm)中的`RoBERTa-wwm-ext-large`PyTorch版本，[SeanLee97/xmnlp: xmnlp](https://github.com/SeanLee97/xmnlp)中的`xmnlp-onnx-models-v5.zip`存入model文件夹中。 21 | 22 | 下载[Embedding/Chinese-Word-Vectors: 100+ Chinese Word Vectors 上百种预训练中文词向量](https://github.com/Embedding/Chinese-Word-Vectors)中的`sgns.sogounews.bigram-char`存入data文件夹中。 23 | 24 | ## 运行 25 | 26 | ### 数据预处理 27 | 28 | ```shell 29 | python web.py 30 | python construct_test.py 31 | python construct_train.py 32 | ``` 33 | 34 | ### 训练模型+预测 35 | 36 | **注**：请手动调整模型的各种数值以达到最优，本项目中默认数值结果不一定为最优。 37 | 38 | #### Word2Vec+CNN 39 | 40 | ```shell 41 | python cnn_model_sen.py 42 | ``` 43 | 44 | #### Bert+CNN 45 | 46 | ```shell 47 | python bert_model_sen.py 48 | ``` 49 | 50 | #### Bert+LSTM 51 | 52 | ```shell 53 | python lstm_model.py 54 | ``` 55 | 56 | ### 机器学习 57 | 58 | ```shell 59 | python ML_model.py 60 | ``` 61 | 62 | ### 结果处理 63 | 64 | **注**：请手动调节分隔数值以达到最优解，本项目中默认数值结果不一定为最优。 65 | 66 | ```shell 67 | python construct.py 68 | ``` 69 | 70 | ## 分数计算 71 | 72 | ```shell 73 | python calc_AUC.py 74 | ``` 75 | 76 | -------------------------------------------------------------------------------- /bert_model_sen.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | import pandas as pd 3 | import matplotlib.pyplot as plt 4 | import tensorflow as tf 5 | import xmnlp 6 | xmnlp.set_model('./model/xmnlp-onnx-models') 7 | 8 | # 训练轮数 9 | epoch = 17 10 | 11 | X_test=np.load("./data/X_test.npy") 12 | X_train=np.load("./data/X_train.npy") 13 | train_data = pd.read_csv('./data/train.news.csv') 14 | test_data = pd.read_csv('./data/test.feature.csv') 15 | Y_train = train_data['label'] 16 | 17 | 18 | ## 加载Title情感分析 19 | # def SentimentAnalysis(text): 20 | # x = list(xmnlp.sentiment(text)) 21 | # x.extend([0.]*1022) 22 | # return x 23 | # train_sen=np.zeros((10587,1,1024)) 24 | # test_sen=np.zeros((10141,1,1024)) 25 | # for i in range(10587): 26 | # train_sen[i,0,:]=SentimentAnalysis(train_data['Title'][i]) 27 | # for i in range(10141): 28 | # test_sen[i,0,:]=SentimentAnalysis(test_data['Title'][i]) 29 | # X_test=np.concatenate((X_test,test_sen),axis=1) 30 | # X_train=np.concatenate((X_train,train_sen),axis=1) 31 | 32 | 33 | ## 是否排除Content情感分析 34 | # X_test=X_test[:,1:,:] 35 | # X_train=X_train[:,1:,:] 36 | 37 | ## 加载文章是否被删除的标记 38 | # file=open('./data/url_check.txt','r') 39 | # import json 40 | # train_web=json.loads(file.readline()) 41 | # test_web=json.loads(file.readline()) 42 | # file.close() 43 | # train_web_array=np.zeros((10587,1,1024)) 44 | # test_web_array=np.zeros((10141,1,1024)) 45 | # for i in range(10587): 46 | # train_web_array[i,0,0]=float(train_web[str(i+1)]) 47 | # for i in range(10141): 48 | # test_web_array[i,0,0]=float(test_web[str(i+1)]) 49 | # X_test=np.concatenate((X_test,test_web_array),axis=1) 50 | # X_train=np.concatenate((X_train,train_web_array),axis=1) 51 | 52 | # for i in range(10587): 53 | # X_train[i,0,-1]=float(train_web[str(i+1)]) 54 | # for i in range(10141): 55 | # X_test[i,0,-1]=float(test_web[str(i+1)]) 56 | 57 | np.random.seed(1) 58 | np.random.shuffle(X_train) 59 | np.random.seed(1) 60 | np.random.shuffle(Y_train) 61 | print("finish shuffling data") 62 | 63 | split = len(X_train) // 7 64 | X_validation = X_train[:split] 65 | X_train_split = X_train[split:] 66 | Y_validation = Y_train[:split] 67 | Y_train_split = Y_train[split:] 68 | print("finish loading data") 69 | 70 | def cnn(X_train): 71 | model = tf.keras.Sequential([ 72 | tf.keras.layers.Convolution1D(input_shape=(X_train.shape[1], X_train.shape[2]), 73 | filters=128, kernel_size=3, activation='relu'), 74 | tf.keras.layers.Dropout(rate=0.5), 75 | tf.keras.layers.MaxPool1D(), 76 | tf.keras.layers.Convolution1D(128, 4, activation='relu'), 77 | tf.keras.layers.Dropout(rate=0.5), 78 | tf.keras.layers.MaxPool1D(), 79 | tf.keras.layers.Convolution1D(64, 5), 80 | tf.keras.layers.Dropout(rate=0.5), 81 | tf.keras.layers.Flatten(), 82 | tf.keras.layers.Dropout(rate=0.5), 83 | tf.keras.layers.Dense(2, activation='softmax'), 84 | ]) 85 | model.compile(loss='sparse_categorical_crossentropy', 86 | optimizer=tf.keras.optimizers.Adam(), 87 | metrics=['accuracy']) 88 | print(model.summary()) 89 | return model 90 | 91 | 92 | model = cnn(X_train) 93 | 94 | print("finish creating model") 95 | 96 | history = model.fit(X_train_split, Y_train_split, epochs=epoch, 97 | batch_size=128, verbose=1, 98 | validation_data=(X_validation, Y_validation)) 99 | 100 | model.save('fakenews_model') 101 | print("finish training model") 102 | 103 | predictions = model.predict(X_test) 104 | np.savetxt('predict.csv', predictions, delimiter=',', fmt='%f') 105 | print("finish predicting") 106 | 107 | 108 | plt.plot(history.history['accuracy']) 109 | plt.plot(history.history['val_accuracy']) 110 | plt.legend(['training', 'validation'], loc='upper left') 111 | plt.savefig('1.png') -------------------------------------------------------------------------------- /calc_AUC.py: -------------------------------------------------------------------------------- 1 | import pandas as pd 2 | from sklearn.metrics import roc_auc_score 3 | 4 | def calc(rightFileName="./data/right.csv", outputFileName="submit.csv"): 5 | output=pd.read_csv(outputFileName) 6 | right=pd.read_csv(rightFileName) 7 | 8 | output_data=output["label"] 9 | right_data=right["label"] 10 | return roc_auc_score(right_data,output_data) 11 | 12 | 13 | if __name__ == "__main__": 14 | print("aoc:",end='') 15 | print(calc()) -------------------------------------------------------------------------------- /cnn_model_sen.py: -------------------------------------------------------------------------------- 1 | import pandas as pd 2 | import jieba 3 | import numpy as np 4 | import tensorflow as tf 5 | import matplotlib.pyplot as plt 6 | import xmnlp 7 | 8 | # 获取数据 9 | train_data = pd.read_csv('./data/train.news.csv') 10 | test_data = pd.read_csv('./data/test.feature.csv') 11 | print("finish read csv") 12 | 13 | 14 | 15 | # 获取预训练 word2vec 并构建词表 16 | word2vec = open("./data/sgns.sogounews.bigram-char", "r", encoding='UTF-8') 17 | t = word2vec.readline().split() 18 | n, dimension = int(t[0]), int(t[1]) 19 | print(n) 20 | print(dimension) 21 | wordAndVec = word2vec.readlines() 22 | wordAndVec = [i.split() for i in wordAndVec] 23 | vectorsMap = [] 24 | word2index = {} 25 | index2word = {} 26 | for i in range(n): 27 | vectorsMap.append(list(map(float, wordAndVec[i][len(wordAndVec[i]) - dimension:]))) 28 | word2index[wordAndVec[i][0]] = i 29 | index2word[i] = wordAndVec[i][0] 30 | 31 | word2vec.close() 32 | print("finish reading") 33 | 34 | # 情感判断 35 | def SentimentAnalysis(text): 36 | xmnlp.set_model('./model/xmnlp-onnx-models') 37 | x = list(xmnlp.sentiment(text)) 38 | x.extend([0.]*298) 39 | return x 40 | 41 | # jieba 分词与词向量构建 42 | features_train = [] 43 | features_test = [] 44 | for text,comment in zip(train_data['Title'],train_data['Report Content']): 45 | # print(len(SentimentAnalysis(comment)),len([0])) 46 | word_feature = [SentimentAnalysis(comment)] 47 | for word in jieba.cut(text): 48 | if word in word2index: 49 | word_feature.append(vectorsMap[word2index[word]]) 50 | features_train.append(word_feature) 51 | 52 | for text,comment in zip(test_data['Title'],test_data['Report Content']): 53 | word_feature = [SentimentAnalysis(comment)] 54 | # word_feature=[] 55 | for word in jieba.cut(text): 56 | if word in word2index: 57 | word_feature.append(vectorsMap[word2index[word]]) 58 | features_test.append(word_feature) 59 | 60 | 61 | print("finish creating features") 62 | 63 | 64 | # 模型输入构建 65 | max_len1 = max([len(i) for i in features_train]) 66 | max_len2 = max([len(i) for i in features_test]) 67 | max_len = max(max_len1, max_len2) 68 | X_train = [] 69 | X_test = [] 70 | for sen in features_train: 71 | tl = sen 72 | tl += [[0] * 300] * (max_len - len(tl)) 73 | X_train.append(tl) 74 | for sen in features_test: 75 | tl = sen 76 | tl += [[0] * 300] * (max_len - len(tl)) 77 | X_test.append(tl) 78 | 79 | print("finish creating X_train X_test") 80 | 81 | Y_train = train_data['label'] 82 | 83 | X_train = np.array(X_train) 84 | X_test = np.array(X_test) 85 | Y_train = np.array(Y_train) 86 | 87 | np.random.seed(1) 88 | np.random.shuffle(X_train) 89 | np.random.seed(1) 90 | np.random.shuffle(Y_train) 91 | 92 | split = len(X_train) // 7 93 | X_validation = X_train[:split] 94 | X_train_split = X_train[split:] 95 | Y_validation = Y_train[:split] 96 | Y_train_split = Y_train[split:] 97 | 98 | print("finish loading data") 99 | 100 | 101 | 102 | # 模型构建 103 | def cnn(X_train): 104 | model = tf.keras.Sequential([ 105 | tf.keras.layers.Convolution1D(input_shape=(X_train.shape[1], X_train.shape[2]), 106 | filters=128, kernel_size=3, activation='relu'), 107 | tf.keras.layers.Dropout(rate=0.5), # Added dropout layer after the first convolutional layer 108 | tf.keras.layers.MaxPool1D(), 109 | tf.keras.layers.Convolution1D(128, 4, activation='relu'), 110 | tf.keras.layers.Dropout(rate=0.5), # Added dropout layer after the second convolutional layer 111 | tf.keras.layers.MaxPool1D(), 112 | tf.keras.layers.Convolution1D(64, 5), 113 | tf.keras.layers.Dropout(rate=0.5), # Added dropout layer after the third convolutional layer 114 | tf.keras.layers.Flatten(), 115 | tf.keras.layers.Dropout(rate=0.5), # Dropout layer before the dense layer (unchanged) 116 | tf.keras.layers.Dense(2, activation='softmax'), 117 | ]) 118 | # lr_schedule = tf.keras.optimizers.schedules.ExponentialDecay( 119 | # initial_learning_rate=1e-2, 120 | # decay_steps=10000, 121 | # decay_rate=0.9) 122 | # model.compile(loss='sparse_categorical_crossentropy', 123 | # optimizer=tf.keras.optimizers.Adam(learning_rate=lr_schedule), 124 | # metrics=['accuracy']) 125 | model.compile(loss='sparse_categorical_crossentropy', 126 | optimizer=tf.keras.optimizers.Adam(), 127 | metrics=['accuracy']) 128 | print(model.summary()) 129 | return model 130 | 131 | 132 | # 模型训练 133 | model = cnn(X_train) 134 | 135 | print("finish creating model") 136 | 137 | history = model.fit(X_train_split, Y_train_split, epochs=20, 138 | batch_size=128, verbose=1, 139 | validation_data=(X_validation, Y_validation)) 140 | 141 | model.save('fakenews_model') 142 | 143 | predictions = model.predict(X_test) 144 | 145 | np.savetxt('predict.csv', predictions, delimiter=',', fmt='%f') 146 | 147 | 148 | plt.plot(history.history['accuracy']) 149 | plt.plot(history.history['val_accuracy']) 150 | plt.legend(['training', 'validation'], loc='upper left') 151 | plt.savefig('1.png') 152 | -------------------------------------------------------------------------------- /construct.py: -------------------------------------------------------------------------------- 1 | import pandas as pd 2 | import csv 3 | 4 | # 分隔数值 5 | split=0.97 6 | 7 | with open('predict.csv', 'r') as f: 8 | predict = csv.reader(f) 9 | pre = [i[0] for i in predict] 10 | prediction = [int(float(i) < split) for i in pre] 11 | id = [str(i) for i in range(1,10141)] 12 | 13 | series1 = pd.Series(prediction) 14 | series2 = pd.Series(id) 15 | print(1, sum(prediction), 0, len(prediction) - sum(prediction)) 16 | 17 | pd.DataFrame({'id': series2, 'label': series1}).to_csv('submit.csv', index=False) -------------------------------------------------------------------------------- /construct_test.py: -------------------------------------------------------------------------------- 1 | from transformers import BertTokenizer, BertModel 2 | import torch 3 | import numpy as np 4 | import pandas as pd 5 | import xmnlp 6 | 7 | device = torch.device("cuda" if torch.cuda.is_available() else "cpu") 8 | print(torch.cuda.get_device_name()) 9 | test_data = pd.read_csv('./data/test.feature.csv') 10 | print("finish read csv") 11 | 12 | model_name = './model/chinese_roberta_wwm_large_ext_pytorch' 13 | tokenizer = BertTokenizer.from_pretrained(model_name) 14 | model = BertModel.from_pretrained(model_name, config='./model/chinese_roberta_wwm_large_ext_pytorch/bert_config.json') 15 | xmnlp.set_model('./model/xmnlp-onnx-models') 16 | 17 | # 情感判断 18 | def SentimentAnalysis(text): 19 | x = list(xmnlp.sentiment(text)) 20 | x.extend([0.]*1022) 21 | return x 22 | 23 | features_test = [] 24 | data_len=len(test_data['Title']) 25 | data_p=int(data_len/100) 26 | index=0 27 | 28 | model.to(device) 29 | model.eval() 30 | 31 | with torch.no_grad(): 32 | for text,comment in zip(test_data['Title'],test_data['Report Content']): 33 | index += 1 34 | if index % data_p == 0: 35 | if index % (data_p*10)