A Framework for Subset Pruning using REP Tree
Abstract
Basically the count of confirmed cases depends upon number of tests so far carried out.The rate of confirmed cases relies on number of test carried out so far. Based on the figures arrived, the Government organization will impose precautionary measures. In case of patient data analysis for any pandemic disease decision making plays crucial role. If the test reports are inaccurate, underreporting takes place. The reports should be provided on time so as to safeguard the lives of patients. Identification of misclassified examples is a well known problem and is drawing significant attention in heath monitoring units. By and by exactness is the primary concern for evaluation and treatment of the individuals.In order to categorize correct classification labels, two typical algorithms are considered, ZeroR and RIPTree. The ZeroR classifies all the instances with majority of the labels without including any predictors. Since the RIPTree is less prone to error, it provides correct information about misclassified instances.The evaluation metrics can be streamlined with the adaptation of these two algorithms. The two algorithmic outcomes can be cross verified with Repeated Incremental Pruning to produce error reduction (RIPPER) algorithm which classifies true positives exactly. The application of the above algorithms assists in confirming the cases, with varied conditions and datasets.