The Risk Analysis solution utilizes prediction analytics to identify students at risk of on-time graduation in order to get them back on track. The prediction service leverages your system’s own (historical) data as well as the Data Lake data to create predictions.
The predictive models use uniquely designed algorithms to predict the probability of student on-time graduation using prior year and current year student data. The predicted graduation probability is expressed as the percentage chance of on-time graduation.
To translate the probabilities into which students are at risk, they are referenced against low, moderate, and high risk thresholds. Cut scores are generated for all actionable factors as well as the overall probability of on-time graduation. A student's probability of on-time graduation can be seen in conjunction with the cut points for all risk levels and their proximity to the nearest graduation probability cut point. This allows you to quickly identify students who may be close to moving between categories.
Imputations are performed on the prediction data sets to adjust for missing data. To reduce the potential for extreme scores being included in the imputed data, the median imputed score is used for prediction. The Imputed Value Range dashboard filter allows you to focus on students with different levels of imputed data (high, moderate, and low).
The solution dashboards contain metrics to track students for on-time graduation. Once students have been identified as having overall risk, the actionable factors (attendance, behavior, and academics) give additional insight into what may be driving risk and where educators may intervene.