嘗試不寫任何代碼完全交由chatgpt做一個股票預測模型之第二部曲

import yfinance as yf
import pandas as pd
import numpy as np
import datetime as dt
import requests

# 設定要取得的股票代碼和時間範圍
ticker = "AAPL"
start_date = dt.datetime(2016, 1, 1)
end_date = dt.datetime(2021, 12, 31)

# 使用 yfinance 取得股票價格資料
df = yf.download(ticker, start=start_date, end=end_date)

# 新增其他相關數據欄位
# 市場情緒：VIX波動率指數
vix_url = "https://fred.stlouisfed.org/graph/fredgraph.csv?id=VIXCLS"
vix_df = pd.read_csv(vix_url, index_col=0, parse_dates=True)
df = pd.merge(df, vix_df, how='left', left_index=True, right_index=True)

# 政治因素：美國總統就職日期
inauguration_dates = {
    'Trump': '2017-01-20',
    'Biden': '2021-01-20'
}
df['Inauguration'] = np.nan
for president, date in inauguration_dates.items():
    inauguration = pd.to_datetime(date)
    df['Inauguration'][df.index >= inauguration] = president

# 移除缺失值
df = df.dropna()

# 儲存處理後的資料
df.to_csv('stock_data.csv')

import pandas as pd
from sklearn.preprocessing import StandardScaler

# 讀取資料
df = pd.read_csv('stock_data.csv', index_col=0)

# 將 Inauguration 欄位做 label encoding
df['Inauguration'] = df['Inauguration'].map({'Trump': 0, 'Biden': 1})

# 計算收盤價的一階差分
df['Close_diff'] = df['Adj Close'].diff()

# 計算 VIX 的 5 日移動平均值
df['VIX_5MA'] = df['VIXCLS'].rolling(window=5).mean()

# 將數據標準化
scaler = StandardScaler()
df_scaled = pd.DataFrame(scaler.fit_transform(df), index=df.index, columns=df.columns)

# 計算收盤價的未來一天差分
df_scaled['Close_diff_shifted'] = df_scaled['Close_diff'].shift(-1)

# 刪除最後一列缺失值
df_scaled = df_scaled[:-1]

# 儲存處理後的資料
df_scaled.to_csv('processed_data.csv')

import pandas as pd
import numpy as np
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras.layers import Dense, LSTM, Dropout

# 讀取資料
df = pd.read_csv('processed_data.csv', index_col=0)

# 分割訓練集和測試集
train_size = int(len(df) * 0.7)
train_data = df.iloc[:train_size, :]
test_data = df.iloc[train_size:, :]

# 標準化數據
scaler = MinMaxScaler()
train_data_scaled = scaler.fit_transform(train_data)
test_data_scaled = scaler.transform(test_data)

# 設定時間窗口大小
window_size = 30

# 創建 X_train 和 y_train
X_train, y_train = [], []
for i in range(window_size, len(train_data_scaled)):
    X_train.append(train_data_scaled[i-window_size:i, :])
    y_train.append(train_data_scaled[i, -2])
X_train, y_train = np.array(X_train), np.array(y_train)

# 創建 LSTM 模型
model = Sequential()
model.add(LSTM(units=50, return_sequences=True, input_shape=(X_train.shape[1], X_train.shape[2])))
model.add(Dropout(0.2))
model.add(LSTM(units=50, return_sequences=True))
model.add(Dropout(0.2))
model.add(LSTM(units=50))
model.add(Dropout(0.2))
model.add(Dense(units=1))

# 編譯模型
model.compile(optimizer='adam', loss='mean_squared_error')

# 訓練模型
model.fit(X_train, y_train, epochs=50, batch_size=32)

# 預測未來的股價
inputs = df.iloc[len(df)-len(test_data)-window_size:, :].values
inputs = scaler.transform(inputs)
X_test = []
for i in range(window_size, len(inputs)):
    X_test.append(inputs[i-window_size:i, :])
X_test = np.array(X_test)
predicted_price = model.predict(X_test)
predicted_price = scaler.inverse_transform(predicted_price)

# 視覺化結果
import matplotlib.pyplot as plt
plt.plot(test_data['Adj Close'].values, color='red', label='Actual Price')
plt.plot(predicted_price, color='blue', label='Predicted Price')
plt.title('Stock Price Prediction')
plt.xlabel('Time')
plt.ylabel('Stock Price')
plt.legend()
plt.show()

y_train.append(train_data_scaled[i, -2])#不是我要預測的 y 收盤價欄位
for i in range(window_size, len(train_data_scaled)):
    X_train.append(train_data_scaled[i-window_size:i, :])
#這一段也很怪應該改成
for i in range(0, len(inputs)-window_size):
    X_train.append(train_data_scaled[i:i+window_size, :])

import yfinance as yf
import pandas as pd
import numpy as np
import datetime as dt
import requests
from textblob import TextBlob
from sklearn.preprocessing import LabelEncoder

# 設定要取得的股票代碼和時間範圍
ticker = "AAPL"
start_date = dt.datetime(2013, 1, 18)
end_date = dt.datetime(2023, 2, 18)

# 使用 yfinance 取得股票價格資料
df = yf.download(ticker, start=start_date, end=end_date)

# 新增其他相關數據欄位
# 市場情緒：VIX波動率指數
vix_url = "https://fred.stlouisfed.org/graph/fredgraph.csv?id=VIXCLS"
vix_df = pd.read_csv(vix_url, index_col=0, parse_dates=True)
df = pd.merge(df, vix_df, how='left', left_index=True, right_index=True)

# 政治因素：美國總統就職日期
inauguration_dates = {
    'Trump': '2017-01-20',
    'Biden': '2021-01-20'
}
df['Inauguration'] = np.nan
for president, date in inauguration_dates.items():
    inauguration = pd.to_datetime(date)
    df['Inauguration'][df.index >= inauguration] = president
le = LabelEncoder()
df['Inauguration'] = le.fit_transform(df['Inauguration'].astype(str))


# 儲存處理後的資料
df.to_csv('stock_data.csv')
import pandas as pd
from sklearn.preprocessing import StandardScaler

# 讀取資料
df = pd.read_csv('stock_data.csv', index_col=0)

# 計算收盤價的一階差分
df['Close_diff'] = df['Adj Close'].diff()

# 計算 VIX 的 5 日移動平均值
df['VIX_5MA'] = df['VIXCLS'].rolling(window=5).mean()

# 計算收盤價的未來一天差分
df['Close_diff_shifted'] = df['Close_diff'].shift(-1)

# 計算未來一天收盤價
df['Close_Tomorrow'] = df['Close'].shift(-1)

df = df.dropna()
df_processed = df

# 儲存處理後的資料
df_processed.to_csv('processed_data.csv')
import pandas as pd
import numpy as np
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras.layers import Dense, LSTM, Dropout

# 讀取資料
df = pd.read_csv('processed_data.csv', index_col=0)

# 分割訓練集和測試集
train_size = int(len(df) * 0.7)
train_data = df.iloc[:train_size, :]
test_data = df.iloc[train_size:, :]

# 標準化數據
scaler = MinMaxScaler()
train_data_scaled = scaler.fit_transform(train_data)
test_data_scaled = scaler.transform(test_data)

# y各自做特徵縮放
y_max = df['Close_Tomorrow'].max()
y_min = df['Close_Tomorrow'].min()
train_data_scaled[:,-1] = (train_data['Close_Tomorrow'].values-y_min)/(y_max-y_min)
test_data_scaled[:,-1] = (test_data['Close_Tomorrow'].values-y_min)/(y_max-y_min)

def inverse_transform_y(y):
  y *= (y_max-y_min)
  y += y_min
  return y

# 設定時間窗口大小
window_size = 30

# 創建 X_train 和 y_train
X_train, y_train = [], []
for i in range(0,len(train_data_scaled)-window_size):
    X_train.append(train_data_scaled[i:i+window_size, :])
    y_train.append(train_data_scaled[i+window_size, -1])
X_train, y_train = np.array(X_train), np.array(y_train)

# 創建 LSTM 模型
model = Sequential()
model.add(LSTM(units=50, return_sequences=True, input_shape=(X_train.shape[1], X_train.shape[2])))
model.add(Dropout(0.2))
model.add(LSTM(units=50, return_sequences=True))
model.add(Dropout(0.2))
model.add(LSTM(units=50))
model.add(Dense(units=1))

# 編譯模型
model.compile(optimizer='adam', loss='mean_squared_error')

# 訓練模型
model.fit(X_train, y_train, epochs=10, batch_size=32)

# 預測未來的股價
inputs = df.iloc[len(df)-len(test_data)-window_size:, :].values
inputs = scaler.transform(inputs)
X_test = []
for i in range(window_size, len(inputs)):
    X_test.append(inputs[i-window_size:i, :])
X_test = np.array(X_test)

train_predicted_price = inverse_transform_y(model.predict(X_train))
test_predicted_price = inverse_transform_y(model.predict(X_test))

# 視覺化結果
import matplotlib.pyplot as plt
plt.plot(train_data['Close_Tomorrow'].values, color='red', label='Actual Price(train)')
plt.plot(train_predicted_price, color='blue', label='Predicted Price(train)')
plt.title('Stock Price Prediction')
plt.xlabel('Time')
plt.ylabel('Stock Price')
plt.legend()
plt.show()

plt.plot(test_data['Close_Tomorrow'].values, color='red', label='Actual Price(test)')
plt.plot(test_predicted_price, color='blue', label='Predicted Price(test)')
plt.title('Stock Price Prediction')
plt.xlabel('Time')
plt.ylabel('Stock Price')
plt.legend()
plt.show()

# 市場情緒：VIX波動率指數
vix_url = "https://fred.stlouisfed.org/graph/fredgraph.csv?id=VIXCLS"
vix_df = pd.read_csv(vix_url, index_col=0, parse_dates=True)
df = pd.merge(df, vix_df, how='left', left_index=True, right_index=True)