Implementation of a Neural Network Approach for Predicting Sales Profit

Abstract

The term "artificial neural network" (ANN) refers, in the majority of cases, to a piece of computing hardware that is impacted by the process, function, and cognitive growth that are comparable to that of a human brain. It does this by simulating the way neurons in the human brain carry out their functions. It is able to solve complicated and dynamic issues in real time with some realistic probability constraints because it understands, observes, and recognises the patterns in the data. This paper implements the Microsoft neural network in Microsoft SQL Server Management using visual studio on a dataset that is accessible to the public in order to demonstrate the efficacy of neural networks in transforming raw data into an in-depth understanding of the trends in a dataset that is difficult to visualise. The dataset was chosen because it is readily available to the public. In addition to that, the error margin or standard deviation in the value prediction of the method that was carried out on the database is shown in this work. The selected dataset contains a significant volume of bike sale records from throughout Europe. As a result, it may be categorised as "big data," which cannot be resolved by utilising a standard paper record method or a typical data mining algorithm that is incapable of learning. Additionally, optimizers and modifiers that have an influence on the prediction value that is processed by the SQL database are investigated in this study. Traditional techniques of data analysis, on the other hand, are unable to correctly access or interpret data that is both so complicated and so dynamic as to be able to deliver quality forecasting. Neural networks, on the other hand, are able to resolve the information contained in big data into valuable knowledge in the form of predictions.