An Analysis Of Students' Subsidy Based On Periodic Consumption Characteristics

Xin He,Weizhong Qian,Xiaoyu Li,Yi Feng
DOI: https://doi.org/10.1109/icaibd.2018.8396207
2018-01-01
Abstract:This paper takes the student's consumption behavior as the research object, redefine the representation of behavior data, add a large number of time's slices as its features, and combine the weights with the self-fitting random forest and the FP-Growth algorithm to conduct a periodic iteration of the behavior. The top 20 features that have the greatest impact on the bursary are obtained. Taking the weekly consumption amount feature as an example, the trends in the consumption amount and the bursary are studied. Experiments show that there are obvious rules in both the value and time of the weekly consumption amount and the bursary.
What problem does this paper attempt to address?