linear multivariate regression

In this study, the accuracy of daily rain prediction by two methods including DeepESN neural network and Linear multivariate regression has been compared. For this reason, actual data for rain and other affective parameters by daily distance time and thirty years has been received from the Hormozgan climate organ and has been analyzed by two methods. This data is from Bandar Abbas, Minab, and Qeshm cities, and because of short distance of these cities before entering data to DeepESN neural network and linear multivariate regression has been averaged. Demodulation for DeepESN in Matlab and linear multivariate regression in SPSS has been done. At the end of processing, results show that the daily rain forecasting model in DeepESN neural network has higher accuracy compared with the model produced with linear multivariate regression and have received better results for the DeepESN forecasting model by using estimation functions.

In this study, the accuracy of daily rain prediction by two methods including DeepESN neural network and Linear multivariate regression has been compared. For this reason, actual data for rain and other affective parameters by daily distance time and thirty years has been received from the Hormozgan climate organ and has been analyzed by two methods. This data is from Bandar Abbas, Minab, and Qeshm cities, and because of short distance of these cities before entering data to DeepESN neural network and linear multivariate regression has been averaged. Demodulation for DeepESN in Matlab and linear multivariate regression in SPSS has been done. At the end of processing, results show that the daily rain forecasting model in DeepESN neural network has higher accuracy compared with the model produced with linear multivariate regression and have received better results for the DeepESN forecasting model by using estimation functions.

One of a variety of vehicles that plays a significant role in traffic accident in Iran is heavy vehicles. According to available statistics, more than 8% of whole traffic accidents has been created by heavy vehicles which accounting for about 15% of deaths from accidents. As well as according to current statistics, the reason of 70 percent of traffic accidents in Iran are human, so heavy vehicle drivers can be very effective in causing accidents. In this research, the main human errors of heavy vehicle drivers have been identified by determining the impact of each human error on the rate of accidents, then it is determined which error has the most impact on crashes, as well as the relationship between errors and the number of accidents. The statistical population in this study is heavy vehicle drivers and the case study region is the main road of Zanjan-Qazvin in 2018. The final results show that using a mathematical model derived from law non-compliance parameters, psychological problems and driver fatigue can predict the amount of accidents. In the other words, increasing or decreasing to one unit of the parameters can increase or decrease the number of accidents in the case study path.