An analysis and forecasts of online product sales based on BP Neural Network and Pearson Coefficient

We construct two models to derive the optimal plan. We successfully utilize our models to determine the best online sales strategy and identify potentially important design features that would enhance product desirability. We establish the Genetic Algorithm-Optimized BP Neural Network Model to find the relationship between time-based measures within data sets and the reputation of products. After quantifying all the indicators, we take BP Neural Network Model as a fitting platform to fit and forecast the reputation-timeline figure. To degrade the influence because of the slow rate of convergence, we introduce Genetic Algorithm to seek optimal global solutions and optimize BP Neural Network Model. Finally, we arrive at the conclusion that the reputation of products has potential correlations with time-based measures within data sets. We establish the Pearson Correlation Model to get the correlation between indexes. Firstly, we quantified all the indicators. According to the model analysis, we find that specific star ratings do incite more reviews. Besides, we use VADER, an NLP Sentiment Analysis Model, to quantify the sentiment descriptors of the reviews. Finally, we conclude that specific quality descriptors of text-based reviews are strongly associated with star rating level but barely associated with the helpful rating level.