ON ANALYSIS AND EVALUATION OF COCKTAIL PARTY EFFECT ON APPLIED EDUCATIONAL PRACTICE THEORY USING NEURAL NETWORKS MODELING.

This piece of research deals with an interestingly observed interdisciplinary challenging educational issue associated with children\'s learning performance phenomenon in classrooms. more precisely, it addresses an answer for the critically challenging educational question : how any of the students could focus on teachers\' interactive speaking in noisy environmentally overcrowd class?. more precisely. More specifically, it deals with the critical educational issue of What is the effect of the school environment on students considering Cocktail Party Effect (CPE) on learning achievement and its relation to the Educational Practice Theory. Moreover, this work presents a systematic approach for evaluation of an interdisciplinary phenomenal problem of human\'s selectivity auditory scene analysis. Interestingly, in the educational field practice this Cocktail Party Effect results in the Cocktail Party Problem (CPP) dealing with an auditory perception phenomenon. Additionally, it is noticed that introduced proposal for active audition modeling is motivated by analogous active vision processes, such as that observed during Optical Character Recognition (OCR). In nature, observed OCR as well as pattern recognition processes have to be carried out under non ideal environmental learning condition (under effect of noisy data). More particularly, CPP is motivated to be solved by adopting selective auditory attention or equivalently selective hearing which is a type of selective attention and involves the auditory system of the nervous system. Accordingly, children\'s Selective hearing is characterized as the action in which considerable children\'s ability to focus their attention inside classroom on a specific source of teacher\'s spoken wording signals. Commonly, this process experienced as following one speaker (teachers\' speech) in the presence of another noisy overcrowded signals resulting in CPE at classroom. By the end of this paper, some interesting simulation results presented after taking into account the comparative studies of two essential ANN parameters namely : learning rate and gain factor values. Versus varying neurons\' number of the hidden layer associated to self-organized ANN paradigm model. These results revealed the effect of interrelation between various learning rate values against different values of signal to noisy ratio considering student\'s selective responsive attention.