Study on information system infrastructure in Probolinggo by using Self Organizing Map

The infrastructure of information systems becomes one of the main components that support government managements system. This study is addressed to achieve the infrastructure information mapping in District of Probolinggo. The proposed of the research is getting an infrastructure information map of 29 districts administered that become its reliability recommendation. The system information as website is developed for operational purpose. The Self Organizing Map (SOM) as one of neural networks is used. The clustering results by in field observation data show that the infrastructure information system of 29 districts administered are defined by 3 clusters, specifically 3 districts administered have good infrastructure, 20 have fair infrastructure and 6 districts administered have less infrastructure. Furthermore, the clustering by using SOM shows that have 1 good, 22 fair and 5 less, respectively. As a result, the accuracy of both clustering is less than 70 %. That means, SOM still need in field observation data to evaluate the recommendation.