Destiny of Legal Petition: Accept or Reject?: A Machine Learning Approach to Predict the Legal Petition's Initial Decision

Presently, the Indian judiciary system grapples with around 30 million in pending cases, translating to approximately 73000 cases per judge. This backlog introduces biases into the system, compounded by the undue influence of political figures and affluent individuals, jeopardizing the judiciary's independence. Additionally, issues such as inconsistent data across different courts, the absence of fixed case completion deadlines, rising crime rates, and a surge in Public Interest Litigation (PIL) further exacerbate the problem. Despite the establishment of fast-track courts, tribunals, and other specialized entities to expedite case resolution, the incorporation of technology has become indispensable. A rapid and automated decision support system is urgently required to address backlogs and ensure data coherence within the judiciary. This study focuses on predicting the initial decision for petitions, specifically determining whether a petition should proceed to further legal proceedings based on the statements framed by legal practitioners. This work employs the ILDC dataset, comprising approximately 7500 data points. Various supervised machine learning algorithms, including logistic regression, support vector machine, random forest, decision tree, and naive Bayes, were utilized to train the model. The results revealed that the random forest algorithm yielded the highest accuracy of 79% with a 10% test size.