Stock Market Price Prediction

A college project that uses data acquired from the Yahoo Finance API to predict the closing price of a stock in the National Stock Exchange, India. The main aim of the project is to compare the performance of traditional forecasting techniques for stock pricing with advanced AI based models,and determine their viability in real world scenarios.

Some Screenshots:

Prediction performance chart comparing model output with real world data.

Type	Color
Model Prediction	🟧
Real World Data	🟦

Results

Paper	Algorithm	RMSE	MAE
Proposed Model	CNN	0.0838	0.0534
	CNNLSTM	0.0891	0.0575
	LSTM	0.0827	0.0557
	SVR	0.0940	0.0725
Banik et al, 2022*	LR	0.0753	0.0542
	MA	1.4651	1.2103
	XGBR	1.0362	0.8744
	SVR	0.6015	0.5287
	ARIMA	0.5741	0.4647
	ETS	0.5754	0.4631
	Meanf	0.5754	0.4631
	BoxCox	0.5493	0.4368
	LSTM	0.0413	0.0324

Condensed Project Status: Online , working parttime

Goal	Achieved?	Further Improvements
Predict the closing price of a stock	✅	Better Model Optimization
Use realtime stock market data	✅	Use more precise data, maybe different markets as well?
User selectable Models : LSTM, LSTM-CNN, CNN, SVM	✅	Add more models to the program,like Bi-LSTM
Provides future predictions daily	✅	Adding more time periods: weekly, monthly, yearly
Has a user-friendly GUI	✅	Improving upon GUI
Able to view indicators for a given stock	🚧	Building a better layout for the graphs generated

Developer:

Name	Github id
Harshiv Chandra	@chcheetah

References:

 * Banik, Shouvik, Nonita Sharma, Monika Mangla,
Sachi Nandan Mohanty, and S. Shitharth. "LSTM
based decision support system for swing trading in
stock market." (“LSTM based decision support system 
for swing trading in stock market ...”) Knowledge-
Based Systems 239 (2022): 107994.