• Publications
  • Influence
A Tree Projection Algorithm for Generation of Frequent Item Sets
This paper provides an implementation of the tree projection method which is up to one order of magnitude faster than other recent techniques in the literature and has a well-structured data access pattern which provides data locality and reuse of data for multiple levels of the cache.
Depth first generation of long patterns
An algorithm for mining long patterns in databases by using depth first search on a lexicographic tree of itemsets achieves more than one order of magnitude speedup over the recently proposed MaxMiner algorithm.
Fast Convolution using fermat number transforms with applications to digital filtering
The structure of transforms having the convolution property is developed. A particular transform is proposed that is defined on a finite ring of integers with arithmetic carried out modulo Fermat
New algorithms for digital convolution
It is shown how the Chinese Remainder Theorem (CRT) can be used to convert a one-dimensional cyclic convolution to a multi-dimensional convolution which is cyclic in all dimensions. Then, special
A three-dimensional approach to parallel matrix multiplication
The 3D parallel matrix multiplication approach has a factor of P1" less communication than the 20 parallel algorithms and has been implemented on IBM POWERparallelTM SP2 systems and has yielded close to the peak performance of the machine.
Number theory in digital signal processing
  • R. Agarwal
  • Computer Science
    Proceedings of the IEEE
  • 1 October 1980
Evaluating boosting algorithms to classify rare classes: comparison and improvements
The authors evaluate three existing categories of boosting algorithms from the single viewpoint of how they update the example weights in each iteration, and propose enhanced algorithms in two of the categories, and justify their choice of weight updating parameters theoretically.
PNrule: A New Framework for Learning Classifier Models in Data Mining (A Case-Study in Network Intrusion Detection)
Although no single technique is proven to be the best in all situations, techniques that learn rule-based models are especially popular in the domain of data mining, and can be contributed to the easy interpretability of the rules by humans, and competitive performance exhibited by rule- based models in many application domains.
Number theoretic transforms to implement fast digital convolution
Transforms using number theoretic concepts are developed as a method for fast and error-free calculation of finite digital convolution. The transforms are defined on finite fields and rings of
Mining needle in a haystack: classifying rare classes via two-phase rule induction
This paper designs various synthetic data models to identify and analyze the situations in which two state-of-the-art methods, RIPPER and C4.5 rules, either fail to learn a model or learn a very poor model, and learns a model with significantly better recall and precision levels.