Method and system for predicting carbon deposition index for data drift

The invention discloses a carbon deposition index prediction method and system for data drifting, and the method comprises the steps: obtaining the production data of an actual factory within a period of time, carrying out the preprocessing of the production data, forming a historical sample, and keeping the historical sample in a historical database; preprocessing the test data to obtain a test sample; calculating a similarity factor of the historical sample according to the information of the historical database, finding a similar sample of the test sample, constructing a local model, and giving a prediction result of the test sample by using the local model; and when the actual label is available, updating the test data and the actual label data to a historical database. According to the method, real-time prediction of the carbon deposition quality index is realized, and a better prediction effect is achieved during data drift. 本申请公开了一种用于数据漂移的积碳指标的预测方法及系统，其中方法包括：获取一段时间内实际工厂的生产数据，并对其进行预处理，形成历史样本，并保留到历史数据库中；