An Automated Profession Prediction Application using the Horoscope with Deep Learning-Based Strategies

Astrology is a time-honored and conventional kind of divination that continues to capture people's curiosity. There are no widely accepted rules or astrological prediction principles in the modern world. Astrologers would rather provide excellent services to people without making any guarantees about accuracy than create a set of norms that apply to everyone. Meanwhile, by utilising computers' natural ability to learn, machine learning has shown to be incredibly successful at analysing and forecasting results in a wide range of applications. Learners can manage big, noisy, complex datasets with the help of prediction and classification. The goal of this work is to present a scientific method that emphasises general prediction rules while reducing the drawbacks of the conventional methodology. Three classification methods are used to show that astrology is valid: J48, Logistic-R, and Cursed Gaussian Naïve Bayes. With 10, 12, and 14-fold cross-validation, supervised learning is used to compute a number of metrics, the study also includes the True Positive Rates, False Positive Rates, Precision, and F-Measure statistics. Moreover, true positive rate, false positive rate, accuracy, F-measure, PRC, and ROC area are used to calculate the mean weight of the three class labels—doctor, businessman, and professor. In order to determine which classification strategy performs better, the precision of each method is finally computed and compared. The birthday details (date, location, and timing) of one hundred people from various occupations are gathered in order to carry out this study. Thirty businessmen, thirty physicians, and forty professors are represented in the dataset. Software is used to create each person's horoscope, and a datasheet in.csv format is produced for analysis. Next, the dataset is used in the Weka programme to assess different parameters and get each classifier's accuracy.