Raj Kumar Banothu

Skylines are the objects in databases, which are not dominated by any other object in databases and help us in decision making.Most of the existing skyline algorithms assume a dataset is complete in all dimensions over which skylines need to be computed. This seems to be a very restrictive assumption. A more practical case is that of incomplete datasets (i.e tuples missing values in some dimensions). In this project, we present APURNA - a Java library that can be used directly in Java based… Mehr zeigen

Skylines are the objects in databases, which are not dominated by any other object in databases and help us in decision making.Most of the existing skyline algorithms assume a dataset is complete in all dimensions over which skylines need to be computed. This seems to be a very restrictive assumption. A more practical case is that of incomplete datasets (i.e tuples missing values in some dimensions). In this project, we present APURNA - a Java library that can be used directly in Java based projects to find skylines involving incomplete data. Weniger anzeigen

Application functionalities are :
1. Uninstall protection. 2. Incoming call alert. 3.Outgoing call alert 4. Anti theft alarm.
5.Incoming message alert. 6. Lock mobile with SMS. 7. Track mobile on google maps.

For more information please click on Project Title above.

Implemented Cache Coherent B+ tree, a variant of B+ tree, which has reduced the query time compared to B+ tree by making use of unused space in the internal nodes of B+ tree to cache frequently queried leaf node addresses, thus saving node accesses.Cache Coherent B+ tree has outperformed traditional B+ tree where the query size is huge and certain queries are much more popular than others.

For more information please click on Project Title above.

Classified movies(good or not good) using IMDb movie reviews. Built a system for binary sentiment classification of IMDb movie reviews, i.e movie classified as good if IMDb rating >= 6 and not good if IMDb rating < 5 using predictive models.

Current detection algorithms work well for spam and phishing because it's easy to detect mass generated email sent to millions of addresses. Targeted Malicious Email(TME), on the other hand, targets single users or small groups in low volumes and goes undetected with current conventional detection methods. We have used "Persistent Threat"​ and "Recipient Oriented"​ features to detect TME. Random Forest based TME filter technique with "Persistent Threat"​ and "Recipient Oriented" features… Mehr zeigen

Current detection algorithms work well for spam and phishing because it's easy to detect mass generated email sent to millions of addresses. Targeted Malicious Email(TME), on the other hand, targets single users or small groups in low volumes and goes undetected with current conventional detection methods. We have used "Persistent Threat"​ and "Recipient Oriented"​ features to detect TME. Random Forest based TME filter technique with "Persistent Threat"​ and "Recipient Oriented" features outperformed SpamAssassin, ClamAV, and the joint SpamAssassin+ClamAV filter.

For more information please click on Project Title above. Weniger anzeigen