Authors - Dhvani Shah, Disha Chandaria, Harshil Shah, Janhavi Patel, Abhijit Joshi Abstract - Non-parametric tests are important in statistical analysis where data fails to meet assumptions like normality or homogeneity of variance. Nevertheless, choosing and implementing the right test is frequently statistical in nature, which is a limitation for most users. Therefore, we suggest an intelligent framework that automates the selection and running of non-parametric tests through machine learning. A well-curated dataset of statistical problems is fed into a Decision Tree Classifier, which is labeled as the best model with 93.4 percent accuracy to identify the most appropriate non-parametric test, such as the Runs Test, Wilcoxon Signed-Rank Test, and Mann-Whitney U Test. After prediction, the identified test is run automatically, and outputs are produced in a well-structured, interpretable manner. This method increases accessibility, enhances statistical analysis efficiency, and opens the way to further application of non-parametric testing in data-driven inquiry.
