Data Mining and Knowledge Engineering

Open Access Open Access  Restricted Access Subscription Access
Open Access Open Access Open Access  Restricted Access Restricted Access Subscription Access

Stock Market Analysis using a combination of Textual Data and Numeric Time-Series


Affiliations
1 Department of Computer Science Engineering, The NorthCap University, Gurugram, India
2 The Northcap University, Gurugram, India
     

   Subscribe/Renew Journal


This study identifies the feasibility of predicting stock market using the combination of Sentiment Analysis and Linear Regression. The factors on which the stock market depends were identified and then a consolidation of these factors was constructed to build up a prediction model. This prediction model incorporated both textual data from news articles published by authentic sources as well as numerical data of various economic identifiers. To analyze the textual data, Naïve Bayes Classifier and Lexicon based Sentiment Analysis was carried out and for numerical values, data analysis was performed using linear regression techniques to obtain optimal results. This model computed MMRE of 0.1561.

Lastly, a combination of identifiers that worked the best for prediction of stock values was computed and therefore, prioritized.


Keywords

Analysis, Stock Prediction, Regression, Economic Identifiers, Close Price.
User
Subscription Login to verify subscription
Notifications
Font Size


  • Stock Market Analysis using a combination of Textual Data and Numeric Time-Series

Abstract Views: 508  |  PDF Views: 0

Authors

Aastha Sharma
Department of Computer Science Engineering, The NorthCap University, Gurugram, India
Aadita Gandhi
Department of Computer Science Engineering, The NorthCap University, Gurugram, India
Prerna Singal
The Northcap University, Gurugram, India

Abstract


This study identifies the feasibility of predicting stock market using the combination of Sentiment Analysis and Linear Regression. The factors on which the stock market depends were identified and then a consolidation of these factors was constructed to build up a prediction model. This prediction model incorporated both textual data from news articles published by authentic sources as well as numerical data of various economic identifiers. To analyze the textual data, Naïve Bayes Classifier and Lexicon based Sentiment Analysis was carried out and for numerical values, data analysis was performed using linear regression techniques to obtain optimal results. This model computed MMRE of 0.1561.

Lastly, a combination of identifiers that worked the best for prediction of stock values was computed and therefore, prioritized.


Keywords


Analysis, Stock Prediction, Regression, Economic Identifiers, Close Price.

References