GA-KELM: Genetic-Algorithm-Improved Kernel Extreme Learning Machine for Traffic Flow Forecasting

A prompt and precise estimation of traffic conditions on the scale a few minutes by analyzing past data is crucial for establishing an effective intelligent management system. Nevertheless, because irregularity nonlinear features flow data, developing prediction model with excellent robustness poses significant obstacle. Therefore, we propose genetic-search-algorithm-improved kernel extreme learning machine, termed GA-KELM, to unleash potential improved accuracy generalization performance. By substituting inner product function, short-term forecasting using machines enhanced. The genetic algorithm evades manual traversal all possible parameters in searching optimal solution. performance GA-KELM evaluated eleven benchmark datasets compared several state-of-the-art models. There are four from A1, A2, A4, A8 highways near ring road Amsterdam, others D1, D2, D3, D4, D5, D6, P, close Heathrow airport M25 expressway. On A8, RMSEs 284.67 vehs/h, 193.83 220.89 163.02 respectively, while MAPEs 11.67%, 9.83%, 11.31%, 12.59%, respectively. results illustrate that obviously superior