Base station energy consumption data completion method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of base station energy consumption analysis, and provides a base station energy consumption data completion method and device, electronic equipment and a storage medium. The base station energy consumption data completion method comprises the following steps: collecting energy consumption data of a plurality of associated base stations; performing first time granularity sorting on the energy consumption data of each base station, and performing second time granularity sorting on the energy consumption data under each first time granularity to determine a data missing position; taking the base station number, the first time granularity and the second time granularity as dimensions, and constructing three-dimensional tensors of the energy consumption data of the plurality of associated base stations; performing tensor decomposition on the three-dimensional tensor based on a Bayesian network to obtain a reconstructed tensor; and obtaining an energy consumption value corresponding to the data missing position according to the reconstruction tensor so as to complement the energy consumption data of the plurality of associated base stations. The method can be suitable for a base station energy consumption data completion scene with a small data size, the high-dimensional tensor of the base station energy consumption data is decomposed based on the Bayesian network, and efficient missing number filling is achieved.

Description

technical field [0001] The present invention relates to the technical field of base station energy consumption analysis, in particular to a method, device, electronic equipment and storage medium for completing base station energy consumption data. Background technique [0002] The integrity of base station energy consumption data plays a key role in data statistics and analysis, and AI (Artificial Intelligence, artificial intelligence) training models (used for prediction, decision-making, etc.). However, due to the failure of the collection equipment, the loss of the transmission process, and the error of the storage process, the energy consumption data of the base station is often missing. [0003] Existing data completion schemes, usually based on deep learning and generative adversarial networks, have the following problems: [0004] Existing solutions require a large amount of effective training data, and because the base station covers different areas and scenes, the...

AI Technical Summary

Problems solved by technology

[0004] Existing solutions require a large amount of effective training data, and because the base station covers different areas and scenes, the energy consumption of different base stations is different, and the energy consumption of base stations is seasonal and greatly affected by temperature. Therefore, an effective base station can Consumption of historical data can only use the data of a specific base station in the past one or two weeks, and there are still some missing numbers, resulting in a serious shortage of data. It is difficult to mine useful features from a small amount of data to complete the complement with the existing scheme;

[0005] In addition, the existing scheme uses a complex neural network structure, which requires more computing resources to train the model, and has high requirements for hardware

Images