Establishment and analysis method of power user portrait based on big data technology

Abstract

The invention discloses a method for establishing and analyzing electric power user portraits based on big data technology, which can obtain historical electricity consumption information, basic attributes, payment information and appeal information of electric power users; determine the collection of classification categories of user portraits, and the influence of classification results The factor set determines the mapping relationship between the influencing factor set and the classification set; the obtained data is randomly selected, and part of it is used as a training sample, and the data is used as a prediction sample; the training sample and the prediction sample are normalized, discretized, and Attribute reduction, determine the set of influencing factors after correction; train the training samples, and use ten-fold cross-validation as the test mode, establish a power user portrait prediction model based on the naive Bayesian classifier, use the prediction model to predict the samples Data classification, mining and analysis to obtain power user portraits. The invention facilitates the forecasting and management of electric power flows.

Description

technical field [0001] The invention relates to a method for establishing and analyzing power user portraits based on big data technology. Background technique [0002] Nowadays, more and more industries are beginning to pay attention to the application of user portraits. However, because different industries have different industry backgrounds, application scenarios and user needs, user portraits in different industries cannot be considered the same. The use of user portraits in the financial and banking industries is because the consumption habits of the younger generation of customers have changed. They do not like to go to financial outlets to do business, but choose to use smart devices for financial consumption, and it is difficult to have a product that satisfies everyone at the same time. demand. The telecommunications industry needs to realize real-time and precise marketing through user portraits, such as traffic packages, phone bill packages, etc., and at the sam...

AI Technical Summary

Problems solved by technology

However, this solution has obvious defects in both business and technology: from a business point of view, the payment behavior is only a part of user portraits and credit ratings, and cannot be equated with it; from a technical point of view, the determination of the K value in the K-means algorithm is The key point is that the clustering effect is extremely sensitive to the K value, noise points, and outliers, and often requires many experiments to determine the optimal K value, which is inefficient; since each attribute has different influences on user portraits, we need to assign different weight, but the K-means algorithm cannot determine the specific weight, making the classification results unconvincing

Images