Upload 2 files

app.py

@@ -0,0 +1,68 @@
+import numpy as np
+import pandas as pd
+import streamlit as st
+import plotly.express as px
+from sklearn.model_selection import train_test_split
+from sklearn import preprocessing
+import yfinance as yf
+from sklearn.linear_model import LinearRegression
+
+# Streamlit app
+st.title('CUSTOM Stock Price Prediction 💰')
+st.write('This model predicts upon trends. It will not perform well in volatile history. setting the time frame "max" is recommended. Your predicted days value can not exceed the time frame days. Have fun!')
+# Input widgets
+stock = st.text_input('Stock tag', value='NVDA')
+daysago = st.text_input('Time frame in days (write "max" for max time)', value='365')
+forecast_out = st.number_input('Predicted days', value=180,min_value=1)
+forecast_col = 'Close'
+
+def prepare_data(df,forecast_col,forecast_out):
+    label = df[forecast_col].shift(-forecast_out) #creating new column called label with the last 5 rows are nan
+    X = np.array(df[[forecast_col]]) #creating the feature array
+    X = preprocessing.scale(X) #processing the feature array
+    X_lately = X[-forecast_out:] #creating the column i want to use later in the predicting method
+    X = X[:-forecast_out] # X that will contain the training and testing
+    label.dropna(inplace=True) #dropping na values
+    y = np.array(label)  # assigning Y
+    X_train, X_test, Y_train, Y_test = train_test_split(X, y, test_size=0.2, random_state=42) #cross validation
+
+    response = [X_train,X_test , Y_train, Y_test , X_lately]
+    return response
+
+# Button to trigger model generation and prediction
+if st.button('Generate'):
+    # Fetch stock data
+    if daysago != 'max':
+        daysago = str(daysago) + 'd'
+
+    ticker = yf.Ticker(stock)
+    data = ticker.history(period=daysago)
+
+    X_train, X_test, Y_train, Y_test , X_lately =prepare_data(data,forecast_col,forecast_out); #calling the method were the cross validation and data preperation is in
+
+    # Model Generation
+    learner = LinearRegression()
+    learner.fit(X_train, Y_train)
+    score = learner.score(X_test, Y_test)
+    forecast = learner.predict(X_lately)
+    #st.write('Used Model:', selected_algorithm)
+    st.write('Accuracy Score:', score)
+
+    #GRAPH
+    # Create a DataFrame with future dates and predicted values
+    future_dates = pd.date_range(start=data.index[-1] + pd.Timedelta(days=1), periods=forecast_out, freq='D')
+    predicted_data = pd.DataFrame({'Date': future_dates, 'Predicted Close': forecast})
+
+    # Concatenate original data and predicted data
+    combined_data = pd.concat([data.rename(columns={'Close': 'Actual Close'}), predicted_data.set_index('Date')], axis=1)
+
+    # Plot original and predicted stock prices
+    fig = px.line(combined_data, x=combined_data.index, y=['Actual Close', 'Predicted Close'], title=f'Predicted {stock} Stock Prices')
+    fig.update_layout(xaxis_title='Date',yaxis_title='Price',legend_title_text='')
+
+    # Set line colors
+    fig.data[1].line.color = 'orange'
+
+    st.plotly_chart(fig)
+
+    st.write('Findings: I tried using pycaret.regression to model, interestingly the r2 score of KNeighborsRegressor() was always the highest, with 96%, but it clearly gave a wrong output. Methods other than Ridge, Lasso and Linear Regression seem to always fail so I decied to stick with Linear Regression despite its 88% acc. score. It is not always correct, if the stock price has lots of ups and downs, it wont be able to give a good estimate as the accuracy score often goes below 50%. The discrepancy can be understood by the disconnection between the lines actual close and predicted close. It needs deep learning to go ruther.')