A Real Time Stock tendency prognostication using Quantopian

Author ORCID Identifier

K. Chandrakala:https://orcid.org/0000-0003-2174-532X

Publication Title

Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020

ISBN

9781728184708

Document Type

Conference Proceeding

Abstract

Stock trends prediction project is about designing a strategy that helps in predicting the stock trends in real time. To achieve it, we have considered Quantopian, one of the world's leading financial marketing strategy analysis as our platform. Quantopian follows the 3-Step approach of Alpha coding, Code Optimization Portfolio. Alpha Code is the main algorithm that harbours, our strategy in stock trend prediction. We consider a single factor that influences our code and gradually come up with two or more factors combination to advance efficiency of optimization strategy. Sentiment factor is used to give an efficiency of 9.3% at the end of back testing. When the combination of sentiment, operation ratio and revenue growth are passed to alpha lens, it is found that the efficiency is raised to 64%, making it much more reliable and stable. The accuracy of multiple alpha factors model attains more than 60%, contrast with earlier prediction algorithms with a single alpha factor have around 9% accuracy increases with 51%.

First Page

1260

Last Page

1267

Publication Date

12-1-2020

DOI

10.1109/ICMLA51294.2020.00198

Keywords

Alpha Lens, Portfolio optimization, Quantopian, Stock, trends

Department

Department of Electrical Engineering

Link to Full Text

Share

COinS