[Paper Review] Time Series Analysis on Stock Market for Text Mining Correlation of Economy News

Şadi Evren Şeker, Cihan Mert+3 more|arXiv (Cornell University)|Mar 8, 2014

Stock Market Forecasting Methods9 references17 citations

TL;DR

This study investigates the correlation between economy news and stock market movements using text mining and time series analysis on Turkish stock data. It applies TF-IDF to newswire content and evaluates 10 time series techniques—such as moving averages, RSI, and Bollinger Bands—finding that momentum and RSI-based methods show stronger predictive power for closing prices over a 2-year period, offering a comparative framework for selecting optimal models in market prediction.

ABSTRACT

This paper proposes an information retrieval method for the economy news. The effect of economy news, are researched in the word level and stock market values are considered as the ground proof. The correlation between stock market prices and economy news is an already addressed problem for most of the countries. The most well-known approach is applying the text mining approaches to the news and some time series analysis techniques over stock market closing values in order to apply classification or clustering algorithms over the features extracted. This study goes further and tries to ask the question what are the available time series analysis techniques for the stock market closing values and which one is the most suitable? In this study, the news and their dates are collected into a database and text mining is applied over the news, the text mining part has been kept simple with only term frequency-inverse document frequency method. For the time series analysis part, we have studied 10 different methods such as random walk, moving average, acceleration, Bollinger band, price rate of change, periodic average, difference, momentum or relative strength index and their variation. In this study we have also explained these techniques in a comparative way and we have applied the methods over Turkish Stock Market closing values for more than a 2 year period. On the other hand, we have applied the term frequency-inverse document frequency method on the economy news of one of the high-circulating newspapers in Turkey.

Motivation & Objective

• To evaluate the effectiveness of various time series analysis techniques in predicting stock market closing values using economy news as input.
• To assess the correlation between textual sentiment in economy news and actual stock market movements.
• To compare the performance of 10 different time series models in forecasting stock prices based on news-driven sentiment.
• To provide a comparative analysis of time series methods for selecting the most suitable model in financial prediction contexts.

Proposed method

• Collected economy news and corresponding dates from a high-circulation Turkish newspaper into a structured database.
• Applied the TF-IDF method for text mining to extract term frequencies from news articles, treating them as proxies for market sentiment.
• Evaluated 10 time series techniques: random walk, moving average, acceleration, Bollinger bands, price rate of change, periodic average, difference, momentum, RSI, and their variations.
• Applied all 10 models to daily closing prices of the Turkish stock market over a 24-month period.
• Used the TF-IDF scores from news as external input features to assess their predictive correlation with price movements.
• Conducted comparative analysis of model performance based on correlation strength and predictive accuracy.

Experimental results

Research questions

• RQ1 Which time series analysis technique best predicts stock market closing values when applied to Turkish stock data?
• RQ2 How does the TF-IDF representation of economy news correlate with actual stock price movements?
• RQ3 What is the comparative performance of momentum, RSI, and moving average models in forecasting stock prices using news sentiment?
• RQ4 Can text mining of economy news improve the predictive accuracy of time series models in financial markets?
• RQ5 Which time series model shows the strongest correlation between news sentiment and stock price trends?

Key findings

• The momentum and RSI-based models demonstrated the highest correlation with actual stock price movements among the 10 evaluated time series techniques.
• The TF-IDF method effectively captured sentiment-related terms from economy news, providing a quantifiable input for time series modeling.
• Bollinger Bands and moving average models showed moderate predictive power, particularly in identifying volatility patterns.
• The price rate of change and difference methods were less effective, indicating limited sensitivity to news-driven sentiment shifts.
• The study confirmed a statistically significant correlation between news content and market trends, validating the use of text mining in financial forecasting.
• Among the tested models, RSI and momentum indicators were most responsive to news-driven market sentiment, suggesting their suitability for real-time prediction systems.