Udit Mehta

As databases become increasingly ubiquitous, one of the relevant problems involves maintaining data in a consistent and error free manner across remote systems. Considering the magnitude of data contained in the diverse existing types of databases, the bottleneck occurs at transmission bandwidth of the network during the synchronized exchange of database information during updates. As opposed to a complete cell by cell comparison for all tuples required for database validation - In this paper… Show more

As databases become increasingly ubiquitous, one of the relevant problems involves maintaining data in a consistent and error free manner across remote systems. Considering the magnitude of data contained in the diverse existing types of databases, the bottleneck occurs at transmission bandwidth of the network during the synchronized exchange of database information during updates. As opposed to a complete cell by cell comparison for all tuples required for database validation - In this paper, the proposed design achieves validation of a source database with a remote target in constant space complexity by passing a data structure, Bloom Filter as a message. Any discrepancy is reported in a verbose manner and the false probability rate based on the bloom filter parameters can be tuned to achieve an optimal balance between spurious hits and transmitted message size. The analysis of three bloom filter variants namely Scalable, Cuckoo and Standard Bloom filter is carried out to test feasibility in the context of the proposed design by measuring their computation time and the size of the bloom filter array produced that will be transmitted across the network.
Show less

Proposed a design for a system that uses a combination of the Apriori algorithm for association rule mining and a Naive Bayes classifier for sentiment analysis to make dynamic traffic predictions in Indian metropolitan cities based on Twitter and Facebook social media data after carrying out pre-processing.
Received recognition in the form of a Best Research Poster award at IEEE International Advance
Computing Conference (IACC) 2015 where the future scope of the project such as… Show more

Proposed a design for a system that uses a combination of the Apriori algorithm for association rule mining and a Naive Bayes classifier for sentiment analysis to make dynamic traffic predictions in Indian metropolitan cities based on Twitter and Facebook social media data after carrying out pre-processing.
Received recognition in the form of a Best Research Poster award at IEEE International Advance
Computing Conference (IACC) 2015 where the future scope of the project such as accommodating
local languages, weather and user feedback was presented.

Short video demonstration of project prototype linked in 'See publication' Show less