a study on social security schemes and welfare programmes in unorganised sector with special focus on bidi rollers

Abstract

In this research, the researcher developed a data prediction model that improves the overall newlinenetwork life span and performance in wireless sensor networks employing deep learning. Wireless newlineSensor Networks (WSNs) are network types that do not rely on cables for connectivity and are newlinerather based on either local wireless networks or, more commonly, the internet. Network newlineperformance refers to the response of a network to data, how efficiently it handles the incoming newlinedata, and the amount of data it can efficiently and quickly tackle. Data prediction is the forecasting newlineof the probability of the occurrence of an event keeping the available data as inputs. This research newlineemploys deep learning, an AI-based machine learning process that replicates the learning newlinemechanism of the human brain, to improve the network performance of a WSN. The entire process newlinewas divided into three stages. A deep learning-based algorithm was applied to accumulate data newlinethrough sensor nodes at the first stage. Star topology was put to use for this first phase. A feedforward newlinefilter was employed at the second stage to predict data and compute errors. Finally, an newlineLMS-based model was employed for verification, and data prediction was at last carried out using newlineWNS. Results indicated that data rushed at an average of 3.5 times more speed during the first newlinestage than in a static or any other topological pattern. At the second stage, the application of CNN newlineand LSTM fastened the process of error computation and data prediction by at least 2.7 times. At newlinethe third stage, the overall network performance improved by over 1.5 times. The proposed model newlinesuccessfully improved network performance in a wireless sensor network. newlineKeywords: Wireless sensor network, Convolution Neural Network, Long short-term memory, newline