Privacy-Preserving Knowledge Discovery in Distributed Databases

Data mining is worried with the mining of beneficial information since several kinds of information. In data that concern with human or other specific fields confidentiality anxieties are occupied additional extremely than further data mining responsibilities. In this study vertically, divider the information and mine these separated data at manifold sites. Then, the deep auto-encoders used to classify the yield data, which first auto-encoder learn important features from the input data and the output become input to the auto-encoder2. Furthermore, the auto-encoder2 also reduce the number of features and extracted only the affective features according. Then, the productivity of auto-encoder2 wired to the SoftMax and trained by using supervised technique to classify the features to the classes. Finally, the auto-encoders and SoftMax set and trained in supervised method using labelled data. Several datasets used to validate the proposed method and the obtained results compared with well-known researches in this filed.Data mining is worried with the mining of beneficial information since several kinds of information. In data that concern with human or other specific fields confidentiality anxieties are occupied additional extremely than further data mining responsibilities. In this study vertically, divider the information and mine these separated data at manifold sites. Then, the deep auto-encoders used to classify the yield data, which first auto-encoder learn important features from the input data and the output become input to the auto-encoder2. Furthermore, the auto-encoder2 also reduce the number of features and extracted only the affective features according. Then, the productivity of auto-encoder2 wired to the SoftMax and trained by using supervised technique to classify the features to the classes. Finally, the auto-encoders and SoftMax set and trained in supervised method using labelled data. Several datasets used to validate the proposed method and the obtained results compared with well-known researches in this filed.