Question-5: Machine Learning

Title

A Comparative Study of Pattern Mining-Based Models and Machine Learning Algorithms

Abstract

Pattern mining is an essential approach for identifying significant trends in historical stock data. Candlestick analysis, a widely-used technique, encodes daily stock price movements—open, high, low, and close—into a visual representation. Using identified patterns, subsequence- and similarity-based models can be developed to predict future stock prices.

This project aims to compare the performance of subsequence- and similarity-based models with machine learning algorithms, including tree-based methods, support vector machines, and neural networks, to evaluate their effectiveness in stock price prediction.n.