Diagnosing Mathematical Procedures in Measurement and Geometry of 7th Grade Students Through Machine Learning Platform for Demonstration School of Khon Kaen University, Thailand, Secondary Section (Mor Din Daeng)

Diagnosing the mathematical procedures levels in measurement and geometry content of 7th grade students by using a machine learning platform is to bring the diagnostic results to individual student and overall development appropriate for learners in classrooms. This research aims to diagnose the level of mathematical procedure in measurement and geometry content of 7th grade students by using a machine learning platform both overall and individually, with 32 students, by using design research methodology. This research is in the phase of using a machine learning platform in mathematics classrooms for diagnosing the level of mathematical procedures. This research has brought a data set from the research project on assessment for learning mathematical proficiency to machine learning analysis to predict and classify the level of mathematical procedures. This tool was developed to classify the levels under the Construct Modeling as a basis for tool construction. It has the characteristic of being 7 open-ended situational tests, with item quality meeting the specified criteria. The difficulty values according to Item Response Theory (Rasch model) range from -2.34 to 1.71, with an overall reliability coefficient of Cronbach's alpha of 0.81 and a reliability according to item response theory of 0. 89, with a discrimination power of 1. The analysis and interpretation of data using the Rasch model along with consideration on the Wright Map showed that 2 students (6.25%) were at the medium level, 15 students (46.875%) were at the basic level, and 15 students (46.875%) were at the starter level. No students reached the high and advanced levels.