Predecitve model for Stock Return forecast (future prediction) for FTS100 Tech-Mark Series (top technical firms) in UK listed on London Stock Exchange

import pandas as pd import sklearn as sl import numpy as np import matplotlib.pyplot as plt import statsmodels.api as sm from statsmodels.tsa.arima.model import ARIMA from statsmodels.graphics.tsaplots import plot_acf, plot_pacf from statsmodels.tsa.stattools import adfuller from statsmodels.tsa.seasonal import seasonal_decompose from statsmodels.tsa.holtwinters import ExponentialSmoothing from statsmodels.tsa.statespace.sarimax import SARIMAX from statsmodels.tools.eval_measures import rmse from arch import arch_model

Install and import required libraries

import yfinance as yf

Download stock data

data = yf.download("SN.L", start="2015-01-01", end="2022-12-31")

Extract close price

CLOSEPRICE = data["Close"].dropna()

Summary statistics

print(CLOSEPRICE.describe())

Plot close price

plt.figure(figsize=(10, 6)) plt.plot(CLOSEPRICE) plt.title("Stock Close Price, 2015-2022") plt.xlabel("Days") plt.ylabel("Price") plt.show()

Train-test split

train = CLOSEPRICE[:1518] test = CLOSEPRICE[1518:]

Plot train and test data

plt.figure(figsize=(10, 6)) plt.plot(CLOSEPRICE, label="Close Price") plt.plot(train, label="Training Data", color="blue") plt.plot(test, label="Testing Data", color="green") plt.legend(loc="lower right") plt.title("Stock Close Price, 2015-2022") plt.xlabel("Days") plt.ylabel("Price") plt.show()

Augmented Dickey-Fuller test

adf_result = adfuller(train) print("ADF Statistic:", adf_result[0]) print("p-value:", adf_result[1])

Autocorrelation and Partial Autocorrelation plots

fig, ax = plt.subplots(2, 1, figsize=(10, 8)) plot_acf(train, ax=ax[0]) plot_pacf(train, ax=ax[1]) plt.show()

Differencing

diff_train = train.diff().dropna()

Augmented Dickey-Fuller test after differencing

adf_result_diff = adfuller(diff_train) print("ADF Statistic (After Differencing):", adf_result_diff[0]) print("p-value (After Differencing):", adf_result_diff[1])

Autocorrelation and Partial Autocorrelation plots after differencing

fig, ax = plt.subplots(2, 1, figsize=(10, 8)) plot_acf(diff_train, ax=ax[0]) plot_pacf(diff_train, ax=ax[1]) plt.show()

ARIMA model

model = ARIMA(train, order=(2, 1, 2)) model_fit = model.fit() print(model_fit.summary())

Residual analysis

residuals = model_fit.resid fig, ax = plt.subplots(2, 1, figsize=(10, 8)) plot_acf(residuals, ax=ax[0]) plot_pacf(residuals, ax=ax[1]) plt.show()

Forecasting

forecast = model_fit.forecast(steps=503) forecast_values = forecast[0]

Plot forecast

plt.figure(figsize=(10, 6)) plt.plot(forecast_values, label="Forecasted Values") plt.title("Forecasted Stock Close Price") plt.xlabel("Days") plt.ylabel("Price") plt.legend() plt.show()

Accuracy evaluation

rmse_value = rmse(test, forecast_values) print("RMSE:", rmse_value)

GARCH model

model_garch = arch_model(returnss, vol='Garch', p=1, q=1) model_garch_fit = model_garch.fit() print(model_garch_fit.summary())

Forecasting using GARCH model

forecast_garch = model_garch_fit.forecast(start=len(returnss), horizon=503) forecast_values_garch = forecast_garch.variance.values[-1, :]

Plot forecast using GARCH model

plt.figure(figsize=(10, 6)) plt.plot(forecast_values_garch, label="Forecasted Values (GARCH)") plt.title("Forecasted Stock Close Price (GARCH)") plt.xlabel("Days") plt.ylabel("Price") plt.legend() plt.show()

Accuracy evaluation using GARCH model

rmse_value_garch = rmse(test, forecast_values_garch) print("RMSE (GARCH):", rmse_value_garch)