Using Natural Language Processing to Optimize Behavioral Data Analysis in Special Education

Abstract

Special education teachers track the behavior of their students so they can perform a Functional Behavioral Assessment which can ultimately lead to a Behavioral Intervention Plan. These plans can help select what student-specific strategies are best suited for intervention with oppositional behavior. Current methods of tracking can be time-consuming, costly, and difficult to analyze. A natural language processing algorithm has been developed which demonstrates the ability to process text inputs, identify the key topics surrounding oppositional behavior, and assign a positive, negative, or neutral sentiment to the topic. The prototype software is accurate more than 50% of the time.