An Ensemble Based Astrological Prediction Model for Profession and Marriage Using Machine Learning Strategies

Abstract

The fascination with astrology, an ancient and conventional form of prediction, continues to grow despite the absence of universal astrological prediction rules or principles globally. While accuracy is not guaranteed, astrologers prioritize offering high-quality services over establishing universal standards. In contrast, machine learning yields superior outcomes across diverse applications through its capacity to handle large, noisy, complex datasets via classification and prediction. This paper aims to present a scientific method that addresses the shortcomings of traditional astrology, identifies universal prediction rules, and employs classification techniques—Neural Network (NN), Import Vector Machine (IVM), Random Forest (RF), and Iterative Boosting—to validate the reliability of astrology in predicting profession and marriage outcomes. We computed Correctly Classified Instances (CCI), Incorrectly Classified Instances (ICI), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Relative Absolute Error (RAE) using cross-validation with 10, 12, and 14 folds. Additionally, we evaluated F-Measure, Precision, True Positive Rate, False Positive Rate, and area values for MCC, ROC, and PRC. For three-class labeling of professor, businessman, and doctor, we determined the true positive rates, false positive rates, accuracy, F-measure, PRC, and ROC area. We gathered birthdate, birthplace, and time of birth data from one hundred individuals across these professions, creating horoscopes using software. Data analysis involved building a datasheet in .csv format and employing the Weka tool to assess various parameters, including classifier accuracy, to identify the most effective classification method.