Developing Dropout Predictive System for Secondary Schools Using Classification Algorithm

Recently, there has been an increase of enrollment rate in government schools, as a result of fee free and expansion of compulsory basic education to form four in Tanzania. However, the completion rate of students is highly affected by extreme dropout rate. Researchers in previous studies have explored the causes of school dropout, and they came with general recommendation based on treatment measures. This study, however, deals with predictive measures in which classification algorithm is used in developing dropout predictive system. The targeted population of this study was obtained by employing purposive and non-probability sampling techniques. The study was guided by system theory and conducted in four councils of Tabora region in Tanzania because of high rate school dropout reported in the previous studies. After the analysis, it has been observed that social factors and academic factors have strong impact on the targeted variable dropout time. The study recommends the use of dropout predictive system in secondary schools so as to predict future outcomes of students earlier.