Extreme Gradient Boosting
for Predicting Stock Price Direction
in Context of Indian Equity Markets
Sachin Jadhav , Vrushal Chaudhari , Pratik Barhate ,
Kunal Deshmukh , and Tarun Agrawal
Abstract Algorithmic Trading is becoming ever increasing field. We implemented
various machine learning methods to predict stock price direction in the following
paper. Technical Indicators are taken as features and models are trained upon them.
Stock price direction forecasting helps us to choose from daily trading techniques and
increases the probability of having massive profits while maintaining low-risk profile.
In our research extreme gradient boosting (XGBoost) technique has given the highest
accuracy, i.e. 73.1% and have outperformed other machine learning techniques used.
Our Proposed model outperforms existing models in literature and adds forecasting
on day-by-day basis.
Keywords Trading · XGBoost · Equity markets · Algorithmic model · Machine
learning for finance · NSE · NIFTY50
1 Introduction
Predicting the stock price direction is an ongoing topic of discussion among all
financial forums. The main motive behind this trend to get profits from stock market
while minimizing the risks associated. It can be achieved through algorithmic trading,
using machines instead of humans to do the actual trades. Stock price direction
prediction can be done in two ways, i.e. long-term stock price prediction and short-
term stock price prediction. While long-term stock price prediction is usually done
through fundamental analysis of stock, Short-term stock price prediction is done
using technical indicators. Various new techniques like Data mining, Sentimental
analysis and Arbitrage analysis are also being used by professionals to achieve desired
results.
In this paper, we will be focussing on short-term prediction of stock price direc-
tion. Techniques such as sma-rsa overlap, morning breakout, start gap analysis,
momentum-price and market modules are used in Technical Analysis. But the main
S. Jadhav (B ) · V. Chaudhari · P. Barhate · K. Deshmukh · T. Agrawal
Department of Information Technology, PCCoE, Pune, Maharashtra, India
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022
A. K. Nagar et al. (eds.), Intelligent Sustainable Systems,
Lecture Notes in Networks and Systems 334,
https://doi.org/10.1007/978-981-16-6369-7_28
321