ICARE™ Predictive Analytics

Immediate Case Accountability and Risk Escalation

This brief introduction answers some basic questions about ICARE. Please download the attached PDF for a more in-depth understanding of Mindshare's Predictive Analytics module.

Mindshare's Predictive Analytics module analyzes facts within data sets to determine the probability of a certain outcome to an individual child.

Prediction: Reunification Failures

Training Dataset

Using three years of data, known as the training dataset, and a combination of data science and algorithms, this module analyzes the fact patterns for successful reunifications and for failed reunifications.

Current Cohort of Children Reunified (12 months)

Once the model is complete, the model will be loaded with data from a current cohort of children reunified. For example, this can be the most recent 12 months to allow the quality assurance team to see the complete spectrum of probabilities across the most recently reunified to the children nearing their 12 month mark. 

"Probability of Failure" Prediction

The prediction is displayed on the landing page of this dashboard. The prediction, "probablity of failure" is a listing that shows the probability of a poor outcome, by child and case.  

Performance Stratification: Attribute Ranking 

Click on Performance Stratification to access an atrribute ranking showing each data attribute in weight order. Attributes with the highest impact on the model at the top are displayed at the top and in descending order. 

Attached files: PA_MODEL_FAILED_REUNIFICATIONS.pdf