DETECTION OF CONFIRMATION BIAS IN HOROSCOPE TEXTS USING RANDOM FOREST CLASSIFIER

The Random Forest algorithm, a widely used machine learning technique, is utilized in this research to identify confirmation bias in horoscope text. A common feature of horoscopes, confirmation bias is the tendency to interpret vague or general information in a way that reinforces already established beliefs. Horoscope excerpts having labeled dataset was analyzed using a combination of TF-IDF vectorization, n-gram analysis, linguistic markers associated with bias like vague terminology and feature oriented language. The model’s effectiveness was evaluated using the standard metrics such as accuracy, precision, recall and F1 score. Performance of Random Forest was evaluated and compared to a baseline classifier model using Logistic Regression. Results showed that the proposed Random Forest model clearly outperformed the baseline model Logistic Regression, achieving an accuracy of 98.74%, precision of 91%, recall of 72%, and F1-score of 80% in distinguishing biased from unbiased texts. These findings demonstrated the efficiency of machine learning models to detect confirmation bias in text and contribute to advancing both computational linguistics and behavioral psychologyThe research also examined the difficulties associated with subjective annotation in detecting bias, emphasizing the difficulty of defining and classifying confirmation bias in written material. The ability to handle large amounts of data and thereby to provides information about the features to help to identify biases is high using Random Forest algorithm compared to the baseline model Logistic Regression. Machine learning can be utilized to detect psychological biases in text, particularly in fields where language is deliberately ambiguous and susceptible to interpretation. Based on the findings, it is very evident that these techniques could be associated with a variety of tasks includes study on Natural Language Processing, sentiment analysis, and misinformation identification. It advances the disciplines of behavioral psychology and computational linguistics by offering a methodical way to comprehend cognitive biases through automated text analysis.