Modelling the Odds of Occurrence of Positive Narrow Bipolar Pulses Based on the Wave Characteristics

Statistical association of five wave characteristics of Narrow Bipolar Pulses (NBP) were investigated with respect to the two types of NBP, which were Positive Narrow Bipolar Pulse (PNBP) and Negative Narrow Bipolar Pulse (NNBP). The five wave characteristics were Rise Time (RT), Pulse Duration (PD), Slow Front Duration (SFD), Zero Crossing Time (ZCT) and Full Width at Half Maximum (FWHM). As the first step normality check was done for the data collected under the five variables which resulted in non-normal distribution in all five variables. Thereafter, the result of Mann-Whitney test proved that there is a significant distributional difference in PNBP and NNBP with respect to the five wave characteristics. Furthermore, they were used to fit a binary logistic regression model in order to model the odds of the occurrence of PNBP with respect to the significant variables identified in the Mann-Whitney test. This resulted RT, PD, SFD and ZCT to significantly explain the variability the response variable while FWHM did not significantly explained the same. In addition to that the overall fitted model was able to successfully explain 90.3% of the variability of the response variable. Finally, the change in odds were interpreted in terms of the significant predictor variables of the fitted model.