Abstract:

The process of data mining to research student assessment data is very intriguing in the field of secondary education. In this paper, we have applied the data analytics methods of the Principal Component Analysis (PCA) and hierarchical and flat clustering algorithms to student assessment data to discover unknown patterns, trends, and characteristic associations for predicting and assessing success outcomes that would not otherwise be detected. Clustering of data allows student assessment data to be grouped and enables key policymakers, administrators, and school districts to make more informed, data-driven decisions within their schools. The student data clusters within this study is divided into 5 groups to gain the performance characteristics of students: advanced, proficient, passing, basic, and minimum. This study has utilized student data and four variables from State Assessment courses in English (English 2), History (U.S. History), Mathematics (Algebra 1), and Science (Biology 1) focusing on freshman, sophomore, and junior level high school students in a public school district. The analytics methodology implored here can be adopted by school systems

to assist in informing educators about the importance of using data in assessing student performances on the different metrics of measuring learning progressions and increasing graduation rates in secondary education. The importance of the analytics lies in the fact that they are amenable to Artificial Intelligence (AI) to automate and ease the process by analyzing large and higher dimensional sets of data, spotting of significant patterns and generation of single output that navigates educational policy makers towards a particular decision based on predicted assessment outcomes.

Keywords:

Secondary education, student assessment, clustering algorithms, hierarchical clustering, dendrograms, data analytics, data mining, big data, PCA, student data.