As the utilization of the social media platform Twitter grows among individuals, many users use it for illicit activities including, spreading spam. There are various methods and approaches presented by many researchers to combat spam and spammers. Most of these existing approaches considered a balanced dataset having an equal distribution of classes, which is untrue for datasets prevalent in real-life situations. In this chapter, we used an imbalanced dataset. We implemented various resampling approaches such as random undersampling, random oversampling, and SMOTE for handling class imbalance. In the case of an imbalanced dataset, accuracy cannot be considered a better evaluation metric as there might be a bias in favor of the majority class. Instead, evaluation metrics used here are F1 score, precision, recall, and ROC-AUC value. We implemented six different machine learning classifiers on a resampled dataset. Among these techniques, the random forest classifier works well on a resampled dataset using random oversampling with a better F1 score value of 89.06% and a precision of 91.11%. It also shows better F1 scores for both the “spam” and “nonspam” classes.