Improved SAE-BP time sequence video revenue prediction method

Abstract

The invention discloses an improved SAE-BP time sequence video revenue prediction method. The method comprises the following steps: step 1, obtaining user basic data; 2, performing normalization processing on the collected user basic data to obtain a data matrix; step 3, training a sparse auto-encoder model; step 4, training a video revenue prediction model based on SAE-BP; 5, performing fine adjustment on the pre-trained stack sparse auto-encoder to improve the accuracy of the model; step 6, performing fine adjustment on the pre-trained SAE-BP to improve the accuracy of the model; and step 7, embedding the SAE-BP model obtained by training into a platform system, and performing practical application. The invention provides an improved SAE-BP time sequence video revenue prediction method on the basis of data acquisition of a system platform. The SAE features of the data can be automatically extracted through the SAE algorithm, and the hidden layer of the SAE is used as the input layer of the BP to predict the video revenue, so that the accuracy of the model is improved, and a creative and cultural creative worker can judge the influence of the work of the creative and cultural creative worker.

Description

technical field [0001] The invention relates to the field of video revenue prediction, in particular to an improved SAE-BP time series video revenue prediction method. Background technique [0002] Big data technology originated in the 1990s. The theory matured in 2008 with Google's public publication of two papers "Google File System" and "Simple Data Processing Based on Clusters: MapReduce". After that, this technology was widely used in business, science and technology. , medical care, government, education, economy, transportation, logistics and various fields of society. With the continuous maturity of data theory, the continuous improvement of data acquisition methods, and the continuous improvement of data research tools, it is possible to transfer data analysis work from large companies to small teams. [0003] Therefore, we hope to put the power of science and technology into cultural creation through this project. At present, when cultural creators on various tra...

AI Technical Summary

Problems solved by technology

Furthermore, most of the current traffic platforms have a set of content push methods, and this method will only push the content that is easy to be accepted by the audience first, so that cultural and creative workers are faced with the dilemma of low target audience viscosity and unstable content income.

Especially with the influx of individual cultural and creative workers, there will inevitably be problems such as high content similarity, low style recognition, extremely unstable audience groups, and difficulty in discovering individual IPs.

Images