Distributed incomplete pattern matching using unsupervised weighted bloom filter

Pattern matching in distributed incomplete environment can be achieved by collecting all the distributed data at data centre and then perform pattern matching at data centre afterwards. This method will cause the problem of traffic in the network and the system may leads to bottleneck. The proposed technique solves this problem by performing pattern matching at each base station individually. When data is distributed, local data at each base station is incomplete compare to the global data. We use unsupervised weighted bloom filter algorithm (UWBF) which takes only qualified IDs and their corresponding weights from each base station and perform aggregation and verification at data centre. This method will save cost and prevent bottleneck in network. Unsupervised learning is applied here to perform pattern matching on un-trained patterns. The security in network is achieved by applying encryption and decryption techniques.