rasel-harun commited on
Commit
2a985bb
·
verified ·
1 Parent(s): 2bffdc8

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +26 -40
app.py CHANGED
@@ -5,19 +5,17 @@ import matplotlib.pyplot as plt
5
  import yfinance as yf
6
  from sklearn.metrics import mean_squared_error, mean_absolute_percentage_error
7
  import math, warnings
 
8
 
9
- from statsmodels.tsa.arima.model import ARIMA
10
- from prophet import Prophet
11
- from sklearn.preprocessing import MinMaxScaler
12
- from tensorflow.keras.models import Sequential
13
- from tensorflow.keras.layers import LSTM, Dense, Dropout
14
 
15
  warnings.filterwarnings("ignore")
16
 
17
- # Utility functions
18
  def rmse(a, b):
19
  return math.sqrt(mean_squared_error(a, b))
20
 
 
21
  def download_data(ticker, start="2015-01-01", end=None):
22
  df = yf.download(ticker, start=start, end=end, progress=False)
23
  df = df[['Close']].dropna()
@@ -25,6 +23,7 @@ def download_data(ticker, start="2015-01-01", end=None):
25
  df.rename(columns={'Close': 'close'}, inplace=True)
26
  return df
27
 
 
28
  def forecast_models(ticker="AAPL", test_len=180):
29
  df = download_data(ticker)
30
  series = df['close']
@@ -32,10 +31,9 @@ def forecast_models(ticker="AAPL", test_len=180):
32
 
33
  results = {}
34
 
35
- # ========== ARIMA ==========
36
  try:
37
- arima_model = ARIMA(train, order=(5,1,0))
38
- arima_fit = arima_model.fit()
39
  arima_forecast = arima_fit.forecast(steps=len(test))
40
  arima_forecast = pd.Series(arima_forecast, index=test.index)
41
  results['ARIMA'] = {
@@ -46,14 +44,11 @@ def forecast_models(ticker="AAPL", test_len=180):
46
  except Exception as e:
47
  results['ARIMA'] = {"error": str(e)}
48
 
49
- # ========== Prophet ==========
50
  try:
51
- prophet_df = train.reset_index()
52
- prophet_df.columns = ['ds', 'y']
53
- m = Prophet(daily_seasonality=True, weekly_seasonality=True, yearly_seasonality=True)
54
- m.fit(prophet_df)
55
- future = m.make_future_dataframe(periods=len(test), freq='D')
56
- forecast = m.predict(future)
57
  prophet_forecast = forecast.set_index('ds').reindex(test.index)['yhat'].fillna(method='ffill')
58
  results['Prophet'] = {
59
  "forecast": prophet_forecast,
@@ -63,37 +58,29 @@ def forecast_models(ticker="AAPL", test_len=180):
63
  except Exception as e:
64
  results['Prophet'] = {"error": str(e)}
65
 
66
- # ========== LSTM ==========
67
  try:
68
- scaler = MinMaxScaler(feature_range=(0,1))
69
- scaled_series = scaler.fit_transform(series.values.reshape(-1,1))
 
70
 
71
- def create_windows(data, window_size=60):
 
 
72
  X, y = [], []
73
  for i in range(window_size, len(data)):
74
- X.append(data[i-window_size:i, 0])
75
  y.append(data[i,0])
76
  X, y = np.array(X), np.array(y)
77
- X = X.reshape((X.shape[0], X.shape[1], 1))
78
  return X, y
79
 
80
- window_size = 60
81
- X_all, y_all = create_windows(scaled_series, window_size=window_size)
82
  split_index = len(train) - window_size
83
- X_train, y_train = X_all[:split_index], y_all[:split_index]
84
- X_test, y_test = X_all[split_index:], y_all[split_index:]
85
-
86
- model = Sequential([
87
- LSTM(64, return_sequences=True, input_shape=(window_size,1)),
88
- Dropout(0.2),
89
- LSTM(32),
90
- Dropout(0.2),
91
- Dense(1)
92
- ])
93
- model.compile(optimizer='adam', loss='mse')
94
- model.fit(X_train, y_train, epochs=5, batch_size=32, verbose=0) # smaller epochs for demo
95
-
96
- pred_scaled = model.predict(X_test, verbose=0)
97
  pred = scaler.inverse_transform(pred_scaled)
98
  lstm_forecast = pd.Series(pred.flatten(), index=test.index[:len(pred)])
99
 
@@ -132,8 +119,7 @@ def forecast_models(ticker="AAPL", test_len=180):
132
 
133
  return plot_path, metrics_df
134
 
135
-
136
- # Gradio UI
137
  def run_app(ticker):
138
  plot_path, metrics_df = forecast_models(ticker)
139
  return plot_path, metrics_df
 
5
  import yfinance as yf
6
  from sklearn.metrics import mean_squared_error, mean_absolute_percentage_error
7
  import math, warnings
8
+ import joblib
9
 
10
+ from tensorflow.keras.models import load_model
 
 
 
 
11
 
12
  warnings.filterwarnings("ignore")
13
 
14
+ # Utility function
15
  def rmse(a, b):
16
  return math.sqrt(mean_squared_error(a, b))
17
 
18
+ # Download stock data
19
  def download_data(ticker, start="2015-01-01", end=None):
20
  df = yf.download(ticker, start=start, end=end, progress=False)
21
  df = df[['Close']].dropna()
 
23
  df.rename(columns={'Close': 'close'}, inplace=True)
24
  return df
25
 
26
+ # Forecast using pre-saved models
27
  def forecast_models(ticker="AAPL", test_len=180):
28
  df = download_data(ticker)
29
  series = df['close']
 
31
 
32
  results = {}
33
 
34
+ # ===== Load ARIMA =====
35
  try:
36
+ arima_fit = joblib.load("arima_model.pkl")
 
37
  arima_forecast = arima_fit.forecast(steps=len(test))
38
  arima_forecast = pd.Series(arima_forecast, index=test.index)
39
  results['ARIMA'] = {
 
44
  except Exception as e:
45
  results['ARIMA'] = {"error": str(e)}
46
 
47
+ # ===== Load Prophet =====
48
  try:
49
+ prophet_model = joblib.load("prophet_model.pkl")
50
+ future = prophet_model.make_future_dataframe(periods=len(test), freq='D')
51
+ forecast = prophet_model.predict(future)
 
 
 
52
  prophet_forecast = forecast.set_index('ds').reindex(test.index)['yhat'].fillna(method='ffill')
53
  results['Prophet'] = {
54
  "forecast": prophet_forecast,
 
58
  except Exception as e:
59
  results['Prophet'] = {"error": str(e)}
60
 
61
+ # ===== Load LSTM =====
62
  try:
63
+ lstm_model = load_model("lstm_model.h5")
64
+ scaler = joblib.load("scaler.pkl")
65
+ scaled_series = scaler.transform(series.values.reshape(-1,1))
66
 
67
+ # Create sliding windows
68
+ window_size = 60
69
+ def create_windows(data):
70
  X, y = [], []
71
  for i in range(window_size, len(data)):
72
+ X.append(data[i-window_size:i,0])
73
  y.append(data[i,0])
74
  X, y = np.array(X), np.array(y)
75
+ X = X.reshape((X.shape[0], X.shape[1],1))
76
  return X, y
77
 
78
+ X_all, y_all = create_windows(scaled_series)
 
79
  split_index = len(train) - window_size
80
+ X_test = X_all[split_index:]
81
+ y_test = y_all[split_index:]
82
+
83
+ pred_scaled = lstm_model.predict(X_test, verbose=0)
 
 
 
 
 
 
 
 
 
 
84
  pred = scaler.inverse_transform(pred_scaled)
85
  lstm_forecast = pd.Series(pred.flatten(), index=test.index[:len(pred)])
86
 
 
119
 
120
  return plot_path, metrics_df
121
 
122
+ # ===== Gradio UI =====
 
123
  def run_app(ticker):
124
  plot_path, metrics_df = forecast_models(ticker)
125
  return plot_path, metrics_df