NETFLIX STOCK PRICE DIRECTION PREDICTION USING MACHINE LEARNING

Abstract

Predicting stock price movements is a complex and challenging task due to the highly volatile, non-linear, and non-stationary nature of financial markets. This project focuses on predicting the direction of Netflix (NFLX) stock prices using machine learning techniques, specifically Linear Regression combined with time-series feature engineering. Historical stock data from February 2018 to February 2022 is utilized to construct lag-based features that capture temporal dependencies and trends in the data. The developed model is trained to identify patterns and forecast the future movement of stock prices, achieving an impressive R² score of 0.996, which reflects strong predictive performance. To enhance interpretability, interactive visualizations are created using Plotly, enabling clear analysis of stock trends, trading volume, and the comparison between actual and predicted values. Furthermore, the system demonstrates practical applicability by allowing users to input recent data and obtain predictions for the next day’s stock direction. This project highlights that even simple machine learning models, when combined with effective feature engineering, can deliver accurate short-term forecasts and provide valuable insights for financial decision-making.