Occupant harassment is a typical occurrence influencing many individuals all over the globe. Notwithstanding, a lot less privileged individuals have had un-told difficulties from covetous and gainloving landowners or agents who utilize different strategies to make life uncomfortable for them as occupants. This research, therefore, carries out a comparative analysis of developed machine learning algorithms to foresee the chance of an occupant being harassed and the learning model that performs best. Online dataset about tenant's provocation having eleven attributes was downloaded. Relevant features were extracted using the lasso algorithm while Naive Bayes, Logistic Regression and Support Vector Machine (SVM) was utilized for anticipating whether or not there is harassment. The outcome shows that Logistic Regression performs better than the other classifiers with a precision of 95.4% while Naive Bayes has an accuracy of 94.7%, with SVM having minimal exactness of 62%.