ISSN 2291-2118 Log In Register
LIKE US ON FACEBOOK
FOLLOW US ON TWITTER
Call for Paper - October 2019 Edition
IJRBT solicits original research papers for the December Edition of IJRBT. The date of manuscript submission is open
Read More

Email this article (Login required)

A Novel Method to Improve Model fitting for Stock Market Prediction

Sukriti Jain, Samarth Gupta, Amarjot Singh

About The Authors


Sukriti Jain
Dept of Electronics and Communication Engineering, Ambedkar Institute of Advanced Communication technologies and research, GGSIPU, India

Samarth Gupta
Dept of Mechanical and Production Engineering, IIT Rookee, India

Amarjot Singh
Dept of Engineering, Simon Fraser University, Burnaby, Canada
Canada


Abstract


Forecasting the trends of stock market is of extreme importance and profitable stock market traders and also to the researchers who are always trying to find an analogy to describe the behaviour of stock market. Various data mining techniques have been implemented in the recent past to predict the behaviour of stock market.  Every method tries to fit a model to the training data to predict the future. As it is obvious the accuracy of prediction depends on the model fitting. We propose a linear model for stock market prediction and further elaborate on improving the fit of the model. The proposed model and the correction method are tested on Istanbul stock exchange. The proposed model fits the dataset with an average error of 12% which is corrected by proposed method to average of 6%.

Keywords


Stock market; Gaussian Fitting; Linear Models; Technical indicators; Istanbul Stock Exchange Dataset

Number of downloads: 360

Full Text: PDF HTML

Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

Quick Link

Journal Content

  • Search

Peer Review

IJRBT introduces peer-review from its first Edition onwards. The researchers submitting their papers for publication should review atleast one technical paper from their domain. The manuscript also undergoes mandatory procedural review with IJRBT review and scholar panel.