The major issue related to the protection of private information in online analytical processing system (OLAP), is the privacy concern in the adversarial inference or private information from OLAP query answers. The most previous works on privacy preserving OLAP focuses on a single aggregate function and multiple aggregate functions which deal with both exact disclosure and partial disclosure of the data. It is performed using the combination of the simple aggregate functions which guarantees the level of privacy disclosure as required by the user. The malicious user can exploit the correlation among data to infer sensitive information from a series of seemingly innocuous data accesses based on data dependency, database schema and semantic knowledge, the system constructs a semantic inference modal that represents the possible inference channels and the violation detection system keeps track of the user history when the infer sensitive information exceeds the pre specified threshold the current query will be rejected. The violation detection system combines all users query history and rejects the query. The closeness of the user is calculated based on the amount of the information that flow from one user to another.