In modern manufacturing systems, large amounts of data are being collected. Monitoring sources of data and managing that data through many information systems is a challenging task that results in errors. These reduce data quality, the source of which is often data incompleteness. We filled the missing values with the predictions of a machine learning method called neural network. They proved effective in search of data patterns. From a point of view of accuracy and inputting bias the results are better than those from other frequently used imputation methods.
|
