├── README.md ├── config.py ├── train_stock_lstm.py └── LICENSE /README.md: -------------------------------------------------------------------------------- 1 | # Stock Price Prediction using LSTM 2 | Downloads adjusted daily returns of a configurable date range and set of stocks from Yahoo Finance, concatenates them all into a long sequence, and trains an LSTM to predict future returns based on the sequence of past returns. 3 | 4 | ### Specifics 5 | - Implemented in TensorFlow, adapted from Google's PTB RNN prediction example 6 | - Returns are normalized using standard deviation (lookback configurable). Positive drift should be negligible. 7 | - Train / validation / test sets are organized in chronological order. 8 | 9 | ### Dependencies 10 | - TensorFlow 11 | - pandas_datareader 12 | - numpy 13 | 14 | ### Does it work? 15 | Not really, current results aren't much better than chance. The data might be too noisy for this method, or there might be something wrong in the code or model. 16 | -------------------------------------------------------------------------------- /config.py: -------------------------------------------------------------------------------- 1 | import datetime 2 | 3 | names_file = 'data/dow30.csv' 4 | start = datetime.datetime(1990, 1, 1) 5 | #start = datetime.datetime(2013, 1, 1) # for testing 6 | end = datetime.datetime(2015, 12, 7) 7 | save_file = 'data/dow30_{}_{}.npz'.format(datetime.datetime.strftime(start, "%Y%m%d"), 8 | datetime.datetime.strftime(end, "%Y%m%d")) 9 | normalize_std_len = 50 10 | 11 | class SmallConfig(object): 12 | """Small config.""" 13 | init_scale = 0.1 14 | learning_rate = 0.01 15 | max_grad_norm = 5 16 | num_layers = 2 17 | num_steps = 10 18 | hidden_size = 200 19 | keep_prob = 1.0 20 | lr_decay = 0.5 21 | batch_size = 30 22 | 23 | 24 | class MediumConfig(object): 25 | """Medium config.""" 26 | init_scale = 0.05 27 | learning_rate = 0.01 28 | max_grad_norm = 5 29 | num_layers = 2 30 | num_steps = 10 31 | hidden_size = 650 32 | keep_prob = 0.5 33 | lr_decay = 0.8 34 | batch_size = 30 35 | 36 | 37 | class LargeConfig(object): 38 | """Large config.""" 39 | init_scale = 0.04 40 | learning_rate = 0.01 41 | max_grad_norm = 10 42 | num_layers = 2 43 | num_steps = 10 44 | hidden_size = 1500 45 | keep_prob = 0.35 46 | lr_decay = 1 / 1.15 47 | batch_size = 30 48 | -------------------------------------------------------------------------------- /train_stock_lstm.py: -------------------------------------------------------------------------------- 1 | from __future__ import absolute_import 2 | from __future__ import division 3 | from __future__ import print_function 4 | 5 | import pandas_datareader.data as pdr_data 6 | import numpy as np 7 | import time 8 | import os 9 | import sys 10 | from collections import deque 11 | 12 | import tensorflow as tf 13 | from tensorflow.models.rnn import rnn 14 | from tensorflow.models.rnn import rnn_cell 15 | from tensorflow.models.rnn import seq2seq 16 | 17 | import config as c 18 | 19 | """ 20 | Adapted from Google's PTB word prediction TensorFlow tutorial. 21 | 22 | Copyright 2016 Tencia Lee 23 | 24 | Licensed under the Apache License, Version 2.0 (the "License"); 25 | you may not use this file except in compliance with the License. 26 | You may obtain a copy of the License at 27 | 28 | http://www.apache.org/licenses/LICENSE-2.0 29 | 30 | Unless required by applicable law or agreed to in writing, software 31 | distributed under the License is distributed on an "AS IS" BASIS, 32 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 33 | See the License for the specific language governing permissions and 34 | limitations under the License. 35 | """ 36 | 37 | def get_data(): 38 | ''' 39 | If filename exists, loads data, otherwise downloads and saves data 40 | from Yahoo Finance 41 | Returns: 42 | - a list of arrays of close-to-close percentage returns, normalized by running 43 | stdev calculated over last c.normalize_std_len days 44 | ''' 45 | def download_data(): 46 | from datetime import timedelta, datetime 47 | # find date range for the split train, val, test (0.8, 0.1, 0.1 of total days) 48 | print('Downloading data for dates {} - {}'.format( 49 | datetime.strftime(c.start, "%Y-%m-%d"), 50 | datetime.strftime(c.end, "%Y-%m-%d"))) 51 | split = [0.8, 0.1, 0.1] 52 | cumusplit = [np.sum(split[:i]) for i,s in enumerate(split)] 53 | segment_start_dates = [c.start + timedelta( 54 | days = int((c.end - c.start).days * interv)) for interv in cumusplit][::-1] 55 | stocks_list = map(lambda l: l.strip(), open(c.names_file, 'r').readlines()) 56 | by_stock = dict((s, pdr_data.DataReader(s, 'yahoo', c.start, c.end)) 57 | for s in stocks_list) 58 | seq = [[],[],[]] 59 | for stock in by_stock: 60 | lastAc = -1 61 | daily_returns = deque(maxlen=c.normalize_std_len) 62 | for rec_date in (c.start + timedelta(days=n) for n in xrange((c.end-c.start).days)): 63 | idx = next(i for i,d in enumerate(segment_start_dates) if rec_date >= d) 64 | try: 65 | d = rec_date.strftime("%Y-%m-%d") 66 | ac = by_stock[stock].ix[d]['Adj Close'] 67 | daily_return = (ac - lastAc)/lastAc 68 | if len(daily_returns) == daily_returns.maxlen: 69 | seq[idx].append(daily_return/np.std(daily_returns)) 70 | daily_returns.append(daily_return) 71 | lastAc = ac 72 | except KeyError: 73 | pass 74 | return [np.asarray(dat, dtype=np.float32) for dat in seq][::-1] 75 | 76 | if not os.path.exists(c.save_file): 77 | datasets = download_data() 78 | print('Saving in {}'.format(c.save_file)) 79 | np.savez(c.save_file, *datasets) 80 | else: 81 | with np.load(c.save_file) as file_load: 82 | datasets = [file_load['arr_%d' % i] for i in range(len(file_load.files))] 83 | return datasets 84 | 85 | def seq_iterator(raw_data, batch_size, num_steps): 86 | """ 87 | Iterate on the raw return sequence data. 88 | Args: 89 | - raw_data: array 90 | - batch_size: int, the batch size. 91 | - num_steps: int, the number of unrolls. 92 | Yields: 93 | - Pairs of the batched data, each a matrix of shape [batch_size, num_steps]. 94 | The second element of the tuple is the same data time-shifted to the 95 | right by one. 96 | Raises: 97 | - ValueError: if batch_size or num_steps are too high. 98 | """ 99 | raw_data = np.array(raw_data, dtype=np.float32) 100 | 101 | data_len = len(raw_data) 102 | batch_len = data_len // batch_size 103 | data = np.zeros([batch_size, batch_len], dtype=np.float32) 104 | for i in range(batch_size): 105 | data[i] = raw_data[batch_len * i:batch_len * (i + 1)] 106 | 107 | epoch_size = (batch_len - 1) // num_steps 108 | 109 | if epoch_size == 0: 110 | raise ValueError("epoch_size == 0, decrease batch_size or num_steps") 111 | 112 | for i in range(epoch_size): 113 | x = data[:, i*num_steps:(i+1)*num_steps] 114 | y = data[:, i*num_steps+1:(i+1)*num_steps+1] 115 | yield (x, y) 116 | 117 | class StockLSTM(object): 118 | """ 119 | This model predicts a 1D sequence of real numbers (here representing daily stock adjusted 120 | returns normalized by running fixed-length standard deviation) using an LSTM. 121 | It is regularized using the method in [Zaremba et al 2015] 122 | http://arxiv.org/pdf/1409.2329v5.pdf 123 | """ 124 | def __init__(self, is_training, config): 125 | self.batch_size = batch_size = config.batch_size 126 | self.num_steps = num_steps = config.num_steps 127 | size = config.hidden_size 128 | 129 | self._input_data = tf.placeholder(tf.float32, [batch_size, num_steps]) 130 | self._targets = tf.placeholder(tf.float32, [batch_size, num_steps]) 131 | 132 | lstm_cell = rnn_cell.BasicLSTMCell(size, forget_bias=0.0) 133 | if is_training and config.keep_prob < 1: 134 | lstm_cell = rnn_cell.DropoutWrapper(lstm_cell, output_keep_prob=config.keep_prob) 135 | cell = rnn_cell.MultiRNNCell([lstm_cell] * config.num_layers) 136 | 137 | self._initial_state = cell.zero_state(batch_size, tf.float32) 138 | 139 | iw = tf.get_variable("input_w", [1, size]) 140 | ib = tf.get_variable("input_b", [size]) 141 | inputs = [tf.nn.xw_plus_b(i_, iw, ib) for i_ in tf.split(1, num_steps, self._input_data)] 142 | if is_training and config.keep_prob < 1: 143 | inputs = [tf.nn.dropout(input_, config.keep_prob) for input_ in inputs] 144 | 145 | outputs, states = rnn.rnn(cell, inputs, initial_state=self._initial_state) 146 | rnn_output = tf.reshape(tf.concat(1, outputs), [-1, size]) 147 | 148 | self._output = output = tf.nn.xw_plus_b(rnn_output, 149 | tf.get_variable("out_w", [size, 1]), 150 | tf.get_variable("out_b", [1])) 151 | 152 | self._cost = cost = tf.reduce_mean(tf.square(output - tf.reshape(self._targets, [-1]))) 153 | self._final_state = states[-1] 154 | 155 | if not is_training: 156 | return 157 | 158 | self._lr = tf.Variable(0.0, trainable=False) 159 | tvars = tf.trainable_variables() 160 | grads, _ = tf.clip_by_global_norm(tf.gradients(cost, tvars), config.max_grad_norm) 161 | #optimizer = tf.train.GradientDescentOptimizer(self.lr) 162 | optimizer = tf.train.AdamOptimizer(self.lr) 163 | self._train_op = optimizer.apply_gradients(zip(grads, tvars)) 164 | 165 | def assign_lr(self, session, lr_value): 166 | session.run(tf.assign(self.lr, lr_value)) 167 | 168 | @property 169 | def input_data(self): 170 | return self._input_data 171 | 172 | @property 173 | def targets(self): 174 | return self._targets 175 | 176 | @property 177 | def initial_state(self): 178 | return self._initial_state 179 | 180 | @property 181 | def cost(self): 182 | return self._cost 183 | 184 | @property 185 | def output(self): 186 | return self._output 187 | 188 | @property 189 | def final_state(self): 190 | return self._final_state 191 | 192 | @property 193 | def lr(self): 194 | return self._lr 195 | 196 | @property 197 | def train_op(self): 198 | return self._train_op 199 | 200 | 201 | def main(config_size='small', num_epochs=10): 202 | 203 | def get_config(config_size): 204 | config_size = config_size.lower() 205 | if config_size == 'small': 206 | return c.SmallConfig() 207 | elif config_size == 'medium': 208 | return c.MediumConfig() 209 | elif config_size == 'large': 210 | return c.LargeConfig() 211 | else: 212 | raise ValueError('Unknown config size {} (small, medium, large)'.format(config_size)) 213 | 214 | def run_epoch(session, m, data, eval_op, verbose=False): 215 | """Runs the model on the given data.""" 216 | epoch_size = ((len(data) // m.batch_size) - 1) // m.num_steps 217 | print(epoch_size) 218 | start_time = time.time() 219 | costs = 0.0 220 | iters = 0 221 | state = m.initial_state.eval() 222 | for step, (x, y) in enumerate(seq_iterator(data, m.batch_size, m.num_steps)): 223 | cost, state, _ = session.run([m.cost, m.final_state, eval_op], 224 | {m.input_data: x, m.targets: y, m.initial_state: state}) 225 | costs += cost 226 | iters += m.num_steps 227 | 228 | print_interval = 20 229 | if verbose and epoch_size > print_interval \ 230 | and step % (epoch_size // print_interval) == print_interval: 231 | print("%.3f mse: %.8f speed: %.0f ips" % (step * 1.0 / epoch_size, costs / iters, 232 | iters * m.batch_size / (time.time() - start_time))) 233 | return costs / (iters if iters > 0 else 1) 234 | 235 | with tf.Graph().as_default(), tf.Session() as session: 236 | config = get_config(config_size) 237 | initializer = tf.random_uniform_initializer(-config.init_scale, config.init_scale) 238 | with tf.variable_scope("model", reuse=None, initializer=initializer): 239 | m = StockLSTM(is_training=True, config=config) 240 | with tf.variable_scope("model", reuse=True, initializer=initializer): 241 | mtest = StockLSTM(is_training=False, config=config) 242 | 243 | tf.initialize_all_variables().run() 244 | 245 | train_data, valid_data, test_data = get_data() 246 | 247 | for epoch in xrange(num_epochs): 248 | lr_decay = config.lr_decay ** max(epoch - num_epochs, 0.0) 249 | m.assign_lr(session, config.learning_rate * lr_decay) 250 | cur_lr = session.run(m.lr) 251 | 252 | mse = run_epoch(session, m, train_data, m.train_op, verbose=True) 253 | vmse = run_epoch(session, mtest, valid_data, tf.no_op()) 254 | print("Epoch: %d - learning rate: %.3f - train mse: %.3f - test mse: %.3f" % 255 | (epoch, cur_lr, mse, vmse)) 256 | 257 | tmse = run_epoch(session, mtest, test_data, tf.no_op()) 258 | print("Test mse: %.3f" % tmse) 259 | 260 | if __name__ == '__main__': 261 | # make all arguments of main(...) command line arguments (with type inferred from 262 | # the default value) - this doesn't work on bools so those are strings when 263 | # passed into main. 264 | import argparse, inspect 265 | parser = argparse.ArgumentParser(description='Command line options') 266 | ma = inspect.getargspec(main) 267 | for arg_name,arg_type in zip(ma.args[-len(ma.defaults):],[type(de) for de in ma.defaults]): 268 | parser.add_argument('--{}'.format(arg_name), type=arg_type, dest=arg_name) 269 | args = parser.parse_args(sys.argv[1:]) 270 | main(**{k:v for (k,v) in vars(args).items() if v is not None}) 271 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "{}" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright {yyyy} {name of copyright owner} 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. 202 | --------------------------------------------------------------------------------