Child Maltreatment Forecast Using Bigdata Intelligent Approaches

Child Welfare associations collect large datasets that they are required to process to assess the risk posed to children within their living environment. Current methods for dealing with these large datasets reduce the time caseworkers are able to spend with the children assigned to their care. Within the following research work, methods for obtaining trends in child abuse and neglect datasets are outlined using self-populated datasets. The potential of Machine Learning algorithms in supporting child welfare associations is illustrated through the use of two unsupervised machine learning algorithms. The simple unsupervised learning cycle of the C4.5 algorithm together with the Apriori algorithm assist in ensuring changes in trends can be identified. The results from using Big Data intelligent algorithms indicate that child maltreatment cases can be more efficiently prioritized and handled through Big Data methods. The datasets and machine learning algorithms were run and stored on the Hippo Cluster.