-
Notifications
You must be signed in to change notification settings - Fork 1
/
PrPred.py
149 lines (131 loc) · 4.59 KB
/
PrPred.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
import numpy as np
import matplotlib.pyplot as plt
import pandas
import math
import tensorflow as tf
import keras
import datetime as dt
from datetime import datetime
from getDataFromPoloniex import get_data_from_poloniex
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import LSTM
from sklearn.preprocessing import MinMaxScaler
from sklearn.metrics import mean_squared_error
from pandas.plotting import register_matplotlib_converters
register_matplotlib_converters()
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
np.random.seed(7)
data = pandas.read_csv('data/BTC_ETH.csv')[::-1]
data2 = pandas.read_csv('data/BTC_ETH_real.csv')[::-1]
data2.columns = ['date','high','low','open','close','volume','quoteVolume','weightedAverage']
data3 = pandas.read_csv('data/USDT_BTC.csv')[::-1]
data3.columns = ['date','high','low','open','close','volume','quoteVolume','weightedAverage']
def create_dataset(dataset, look_back=1):
dataX, dataY = [], []
for i in range(len(dataset)-look_back-1):
a = dataset[i:(i+look_back), 0]
dataX.append(a)
dataY.append(dataset[i + look_back, 0])
return np.array(dataX), np.array(dataY)
def split_sequence(sequence, n_steps_in, n_steps_out):
X, Y = list(), list()
for i in range(len(sequence)):
# find the end of this pattern
end_ix = i + n_steps_in
out_end_ix = end_ix + n_steps_out
# check if we are beyond the sequence
if out_end_ix > len(sequence):
break
# gather input and output parts of the pattern
seq_x, seq_y = sequence[i:end_ix], sequence[end_ix:out_end_ix]
X.append(seq_x)
Y.append(seq_y)
return np.asarray(X), np.asarray(Y)
dataset = tf.data.Dataset(data3.close.ewm(span=14).mean())[::-1]
print("REAL: ")
data4 = pandas.read_csv('data/BTC_USDT_predicted.csv')
datapred = pandas.DataFrame(data4.EMA_14)
print("Predicted: ")
scaler = MinMaxScaler(feature_range=(0, 1))
ds = dataset.to_numpy()
ds1 = datapred.to_numpy()
ds1 = ds1.flatten('F')
ds2 = ds.flatten('F')
ds2 = ds2[0:288]
ds3 = np.abs(ds1-ds2)
print(ds3)
percents = list()
for i in range(288):
percents.append(100-(ds3[i]*100)/ds2[i])
percents = np.asarray(percents)
print(percents)
dataset = scaler.fit_transform(dataset)
print("DATASET: ")
print(str(type(dataset)))
ds2 = dataset
print(ds2)
train_size = int(len(dataset))
test_size = len(dataset) - train_size
train, test = dataset[0:train_size,:], dataset[train_size:len(dataset),:]
print(len(train), len(test))
print(str(type(train)))
print("Train: ")
trainF = train.flatten('F')
print(str(type(trainF)))
look_back = 3
trainX, trainY = create_dataset(train, look_back)
print("Create Dataset:")
trXXX, trYYY = split_sequence(trr, look_back, 5)
tarrX, tarrY = split_sequence(trainF, look_back, 288)
trXX, trYY = create_dataset(trr, look_back)
print("Split sequence: ")
trXX = np.reshape(trXX, (trXX.shape[0], trXX.shape[1], 1))
trainX = np.reshape(trainX, (trainX.shape[0], trainX.shape[1], 1))
trXXX = trXXX.reshape(trXXX.shape[0], trXXX.shape[1], 1)
print("Reshape: ")
tarrX = tarrX.reshape(tarrX.shape[0], tarrX.shape[1], 1)
ds2 = ds2.reshape(ds2.shape[0], ds2.shape[1], 1)
trainXX = trainX[3740:len(trainX)]
model = Sequential()
model.add(LSTM(100, activation='relu', return_sequences=True, input_shape=(look_back, 1)))
model.add(LSTM(100, activation='relu'))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adam')
model.fit(trainX, trainY, epochs=20, batch_size=1, verbose=2)
print(trainXX.shape)
print(tarrX.shape)
trainPredict = model.predict(trainXX)
model1 = Sequential()
model1.add(LSTM(100, activation='relu', return_sequences=True, input_shape=(look_back, 1)))
model1.add(LSTM(100, activation='relu'))
model1.add(Dense(288))
model1.compile(loss='mean_squared_error', optimizer='adam')
model1.fit(tarrX, tarrY, epochs=100, verbose=0)
print(np.asarray([[80], [90], [100]]).shape)
resPredict = model1.predict(tarrX)
resProba = model1.predict_proba(tarrX)
resPred = scaler.inverse_transform(resPredict)
print("PREDICTED: ")
print(resPred)
trainPredict = scaler.inverse_transform(trainPredict)
print("PROBABILITY: ")
print(resProba)
X = list()
for i in range(288):
X.append(1596525300+300*i)
resPred = resPred.flatten('F')
print(resPred)
trPred = np.flip(resPred, 0)
trPred = trPred[0:288]
df2 = pandas.DataFrame(trPred)
df2['Date'] = [dt.datetime.fromtimestamp(x) for x in X]
df2.reset_index()
df2.columns = ['EMA_14','Date']
print(df2.Date)
resProba = resProba.flatten('F')
resProba = resProba[0:288]
print(len(resProba), len(df2))
plt.plot(df2.Date, resProba)
plt.show()