Preliminary vertical TEC prediction using neural network: Input data selection and preparation

Total electron content (TEC) is a fundamental and most prevailing ionospheric parameter that leads to Global Positioning System (GPS) error source such as delays, poor signal or lost data. Neural Network (NN) based approaches has proven track record in ionospheric process modeling. In this work, a data preparation method was developed to perform neural network based on VTEC forecast over two stations in Malaysia. GPS Ionospheric Scintillation & TEC Monitor (GISTM) at UKM and LGKW became a part of the feasibility study for the development of data sets as inputs to the NN based TEC prediction model. The study period was selected based on the availability of data, which is from January 2011 to December 2012. The factors that influence VTEC performance are identified and processed accordingly, to be used as input parameter for the VTEC prediction NN model development. The selected parameters are seasonal variation, diurnal variation, and sunspot number which have similar conduct with VTEC for the selected 24 months.