This study analyzed the outcomes of 194 individuals living in rural poverty in Guizhou Province using data obtained from a poverty alleviation information system together with several machine-learning tools. First, four dimensions were abstracted from a multiple factor analysis: family background and work condition, self-development, regional context, and health status and labor capacity. Second, five heterogeneous groups were identified through k-medoids clustering based on the above

more »

... . The results showed that the most significant differences among poverty groups were related to health status and household size. The effectiveness of poverty alleviation programs for the different poverty groups was then evaluated. By employing association rule mining, we showed that education assistance is effective in poverty groups in relatively 'good' condition but not in others, whereas the 'Minimum Living Standard' program failed in all groups, particularly when it was the sole program. Thus, our study indicated that additional targeted alleviation programs should be implemented to address the needs of different poverty groups, especially in terms of education assistance, medical care, and medical prevention programs.