A Decision Tree Model for Public Land Information Systems at The National Land Commission, Kenya.

Abstract

Information Technology is acritical tool to produce accurate, reliable and timely information through the development of Information Systems and to be judged successful, it has to satisfy additional criteria in today’s competitive world. National Land Commission (NLC) faced numerous challenges with a paper-based system in service delivery and it resulted to the introduction of Public Land Information System (PLIS) which would help in service delivery performance. Despite the move the organisation rolled back to paper based system. The study sought to identify factors influencing the success of Information System at the NLC and coming up with a decision tree model that will help improve performance of PLIS. The Integrated Success Model for Evaluating Information System in Public Sector form the basis of this study. The decision tree was used to predict, classify and important factors were represented using a tree. C4.5 decision tree model was created. 10-fold cross validation was employed for testing the model. The model accuracy was 67.8%, the negative and positive predictive value was 0.20 while sensitivity and specificity was 0.5. Factors identified as most important were education level, age, the people who interact with the system and the position of the employees.