Noise modeling using multiple probability distribution functions

Abstract

Many research has been conducted to develop an appropriate statistical model for assessing noise level potential. The key parameter in estimating this potential is the noise from the generator, which is inherently random, making statistical methods essential for accurate estimation. Consequently, noise level probabilities can be analyzed using various probability distributions. Accurately determining the probability distribution of noise level values is crucial for evaluating the noise level potential of the university. However, this paper applies the use of lognormal, Weibull, Nakagami, and gamma distributions to datasets from a specific location in Adamawa State University (ADSU), Mubi. To identify the most suitable distribution, the study employs the Kolmogorov-Smirnov (K-S) test and Anderson-Darling (A-D) test along with graphical representations of the cumulative distribution function (CDF) and probability distribution function (PDF). The (K-S) test was found to be the best model over the (A-D) test. Based on both graphical analysis and computed goodness-of-fit results, the gamma distribution was found as the bestfitting model with a fitness of 0.13805, followed by Nakagami, Weibull, and lognormal with 0.14130, 0.14579, and 0.15709, respectively. Additionally, 62.54% was found to be the probability of exceeding the critical point (PECP).