Web system based on neural networks to improve the tax collection in the District Municipality of La Esperanza in the year 2017

Abstract

In the present investigation a WEB System based on neural networks was developed. The objective was to improve the tax collection in the District Municipality of La Esperanza with the purpose of increasing the average number of monitoring of taxpayers who pay more taxes by sector, increasing the average time in the monitoring of each registered property, increasing the average time in the follow-up of the notifications and, finally, increasing the level of accuracy of the tax collection forecast using neural networks. The population was made up of 33,000 contributors. PHP was chosen as the programming language for the development of the WEB System using MySQL as a database manager, and the ICONIX methodology was used as well. Through the experimental design of pre-test and post-test, the results obtained were positive since the average number of monitoring of taxpayers with the current system was 1.414 times, while with the implemented system it was 3.370 times, increasing by 1.956 the number of times. In the monitoring of properties with the current system was 3.818 times while with the proposed system was 9.655 increasing the average number of registered properties by 5.837. In the follow-up of the notifications with the current system, it was 0.950, while with the proposed system it was 3.992, increasing by 3.042 the average in the effective notifications. The level of accuracy in the tax collection forecast with the current system, it was 0.506 times, while with the proposed system it was 1.518, increasing the average number by 1.012. It was concluded that with the implementation of the system, the tax collection in the District Municipality of La Esperanza was significantly improved.