Accurate forecasting of daily crude oil prices is essential for market participants to make informed investment decisions. This paper introduces a novel hybrid time series forecasting technique to predict daily crude oil prices, leveraging the strengths of regression, time series models, and machine learning techniques. The proposed hybrid model integrates Long Short-Term Memory (LSTM) networks, Prophet, and a combined ARIMA-LSTM approach to enhance forecast accuracy. The methodology starts by preprocessing the crude oil price time series to address common issues such as missing values, variance stabilization, and normalization. The cleaned time series data is then partitioned for training and testing the models. The individual models—LSTM, Prophet, and ARIMA-LSTM—are trained on the historical data, and their performance is compared. The results indicate that the hybrid model outperforms traditional models in terms of prediction accuracy, making it a promising tool for forecasting crude oil prices. This approach highlights the potential of hybrid techniques in improving the reliability of financial time series forecasts.

Keywords: Crude Oil Price Forecasting, Time Series Analysis, Hybrid Model, LSTM, ARIMA, Prophet, Machine Learning, Forecast Accuracy, Financial Predictions, Data Pre-processing.