Predictability of School Administrator’s Job Satisfaction through Hybrid Segmentation-based Prediction Model
Year: 2019
Type: article
Abstract: Today, various researchers have amplified the use of data mining techniques in almost all sectors of society.In this study, the effectiveness of data mining techniques, particularly the use of K-Means segmentation in aid of increasing prediction accuracy when used to job satisfaction dataset is observed.Researches on job satisfaction have long been acclaimed as the prime index of worker's wellbeing as well as a core motivational construct that plays key implications for organizational and administrative policy.Due to the advent of technology, well-established mining and extraction of knowledge from databases are optimized.In this paper, the use of predictive models such as C4.5, DT, Naïve Bayes, OneR, KNN, and JRip algorithms was observed.However, prior to prediction, the variables from the dataset obtained from the 157 school administrator-respondents from the Department of Education, Division of Surigao del Norte, Philippines was segmented using the K-Means algorithm.This study strengthens the notion that a reduced number of variables in the dataset increases the accuracy of a prediction model.Simulation results showed that with the integration of K-Means segmentation, an accuracy of 83.54% using Decision Table algorithm is depicted from the previously identified 82.16% using the DT algorithm alone.
Author: Elvis P. Patulin
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: bronze
OpenAlex ID: W2956100943
