Journal Article

Machine Learning Interface Presentation of SARIMA Forecasting Output Using Visnetwork

Abstract

Machine learning interface presentation of SARIMA forecasting output using visNetwork aimed to designed machine Learning with an advanced coding system for each of the variables under investigation. To achieve this, different nodes were used to denote different characteristics. The development of the interface involved the use of visNetwork machine learning language in R programming environment to showcase forecast as required end result of SARIMA model forecasting. The number of forecasting tables or graphs interface might have been up to four as the forecast were made for 3 years; from year 2024 to 2026 for each of the elements; wind, rainfall, humidity and pressure in station A. The implication of the visNetwork algorithm result is the fact that the interface is machine-ready, it can be used to present forecasts of the variable investigated with little or no adjustment. Another area of justification of machine learning interface is that a single interface algorithm can accommodate as many as possible elements or stations’ forecast without need to increase the number of tables or graphs to display.

Download pdf