Stock trend prediction using news sentiment analysis

📝 Abstract

Efficient Market Hypothesis is the popular theory about stock prediction. With its failure much research has been carried in the area of prediction of stocks. This project is about taking non quantifiable data such as financial news articles about a company and predicting its future stock trend with news sentiment classification. Assuming that news articles have impact on stock market, this is an attempt to study relationship between news and stock trend. To show this, we created three different classification models which depict polarity of news articles being positive or negative. Observations show that RF and SVM perform well in all types of testing. Na"ive Bayes gives good result but not compared to the other two. Experiments are conducted to evaluate various aspects of the proposed model and encouraging results are obtained in all of the experiments. The accuracy of the prediction model is more than 80% and in comparison with news random labeling with 50% of accuracy; the model has increased the accuracy by 30%.

💡 Analysis

Efficient Market Hypothesis is the popular theory about stock prediction. With its failure much research has been carried in the area of prediction of stocks. This project is about taking non quantifiable data such as financial news articles about a company and predicting its future stock trend with news sentiment classification. Assuming that news articles have impact on stock market, this is an attempt to study relationship between news and stock trend. To show this, we created three different classification models which depict polarity of news articles being positive or negative. Observations show that RF and SVM perform well in all types of testing. Na"ive Bayes gives good result but not compared to the other two. Experiments are conducted to evaluate various aspects of the proposed model and encouraging results are obtained in all of the experiments. The accuracy of the prediction model is more than 80% and in comparison with news random labeling with 50% of accuracy; the model has increased the accuracy by 30%.

📄 Content

STOCK TREND PREDICTION USING NEWS SENTIMENT ANALYSIS Kalyani Joshi1, Prof. Bharathi H. N.2, Prof. Jyothi Rao3 1Department of Computer Engineering, KJSCE, Mumbai kalyani.joshi@somaiya.edu

2Department of Computer Engineering, KJSCE, Mumbai hodcomp@somaiya.edu

3Department of Computer Engineering, KJSCE, Mumbai jyothirao@somaiya.edu

ABSTRACT Efficient Market Hypothesis is the popular theory about stock prediction. With its failure much research has been carried in the area of prediction of stocks. This project is about taking non quantifiable data such as financial news articles about a company and predicting its future stock trend with news sentiment classification. Assuming that news articles have impact on stock market, this is an attempt to study relationship between news and stock trend. To show this, we created three different classification models which depict polarity of news articles being positive or negative. Observations show that RF and SVM perform well in all types of testing. Naïve Bayes gives good result but not compared to the other two.
Experiments are conducted to evaluate various aspects of the proposed model and encouraging results are obtained in all of the experiments. The accuracy of the prediction model is more than 80% and in comparison with news random labelling with 50% of accuracy; the model has increased the accuracy by 30%.

KEYWORDS Text Mining, Sentiment analysis, Naive Bayes, Random Forest, SVM, Stock trends

1. INTRODUCTION

In the finance field, stock market and its trends are extremely volatile in nature. It attracts researchers to capture the volatility and predicting its next moves. Investors and market analysts study the market behaviour and plan their buy or sell strategies accordingly. As stock market produces large amount of data every day, it is very difficult for an individual to consider all the current and past information for predicting future trend of a stock. Mainly there are two methods for forecasting market trends. One is Technical analysis and other is Fundamental analysis. Technical analysis considers past price and volume to predict the future trend where as Fundamental analysis On the other hand, Fundamental analysis of a business involves analyzing its financial data to get some insights. The efficacy of both technical and fundamental analysis is disputed by the efficient-market hypothesis which states that stock market prices are essentially unpredictable.

This research follows the Fundamental analysis technique to discover future trend of a stock by considering news articles about a company as prime information and tries to classify news as good (positive) and bad (negative). If the news sentiment is positive, there are more chances that the stock price will go up and if the news sentiment is negative, then stock price may go down.

This research is an attempt to build a model that predicts news polarity which may affect changes in stock trends. In other words, check the impact of news articles on stock prices. We are using supervised machine learning as classification and other text mining techniques to check news polarity. And also be able to classify unknown news, which is not used to build a classifier. Three different classification algorithms are implemented to check and improve classification accuracy. We have taken past three years data from Apple Company as stock price and news articles.

2. LITERATURE SURVEY Stock price trend prediction is an active research area, as more accurate predictions are directly related to more returns in stocks. Therefore, in recent years, significant efforts have been put into developing models that can predict for future trend of a specific stock or overall market. Most of the existing techniques make use of the technical indicators. Some of the researchers showed that there is a strong relationship between news article about a company and its stock prices fluctuations. Following is discussion on previous research on sentiment analysis of text data and different classification techniques. Nagar and Hahsler in their research [1] presented an automated text mining based approach to aggregate news stories from various sources and create a News Corpus. The Corpus is filtered down to relevant sentences and analyzed using Natural Language Processing (NLP) techniques. A sentiment metric, called NewsSentiment, utilizing the count of positive and negative polarity words is proposed as a measure of the sentiment of the overall news corpus. They have used various open source packages and tools to develop the news collection and aggregation engine as well as the sentiment evaluation engine. They also state that the time variation of NewsSentiment shows a very strong correlation with the actual stock price movement. Yu et al [2] present a text mining based framework to determine the sentiment of news articles and illustrate its impact on energy demand.

This content is AI-processed based on ArXiv data.