Comparison of Automated Machine Learning Tools for SMS Spam Message Filtering

Short Message Service (SMS) is a very popular service used for communication by mobile users. However, this can be abused executing illegal activities and influencing security risks. Nowadays, many automatic machine learning (AutoML) tools exist which help domain experts lay users to build high-quality ML models with little or no knowledge. In work, classification performance comparison was conducted between three SMS spam message filtering. These are mljar-supervised AutoML, H2O Tree-based Pipeline Optimization Tool (TPOT) AutoML. Experimental results showed that ensemble achieved the best performance. The Stacked Ensemble model, built using in terms of Log Loss (0.8370), true positive (1088/1116), negative (281/287) metrics. There 19.05\% improvement respect TPOT AutoML 5.56\% satisfactory filtering provides potential application automatically determine model perform