FILTERED BOOK FEATURES WITH ARTIFICIAL NEURAL NETWORK: A CASE STUDY OF THE BAN NONG SAENG SAMAKKEE BOOK OF THE AMSS ++ SYSTEM OFFICE OF KHON KAEN PRIMARY EDUCATION AREA 5

Abstract

Receipt letters were documents submitted to education institutes in primary educational service area to inform assigned educational institutes or relevant educational institutes to perform any transactions without specifying specific educational instituted.  Thus, both relevant and irrelevant receipt letters were delivered to the educational institutes and it considerably affected small schools where there were no administrative teachers. Then, other teachers had to check if there were any receipt letters delivered each day and it wasted their teaching time. For this reason, an artificial neural network was developed to sort out the receipt letters using the data of receipt letters from AMSS++ for 4 January 2016-10 January 2018. When tested with 1,348 subjects of the receipt letters, the accuracy of the back-propagation neural network (BPNN) was at 97.60 percent which represented that the analytical efficiency of this model was at good level.