OPTIMIZATION OF FREQUENT REPRESENTATIVE PATTERN SETS USING NAIVE BAYES TECHNIQUE
Journal Title: International Journal of Engineering Sciences & Research Technology - Year 30, Vol 5, Issue 7
Abstract
Frequent item sets plays a vital part in abundant actions of data mining which continuously endeavor to determine fascinating arrangement through number of databases like association rules, sequences, classifiers based on correlations, episodes, clusters and so on. The time period required to generate periodic item sets enacts a crucial aspect. Various algorithms are developed, taking only time aspect into account. This turns up requirement of determining small number recurrent producing patterns. Here, fundamental problems on mining of periodic item sets as well as pattern sets are elaborated. Frequent pattern mining generally develops a large number of recurrent patterns that depict complex concerns on understanding, viewing and next investigation of derived patterns. This results in determining a small number of classical arrangements of patterns to approximate all other patterns to the best suitable extent. In latest practices in the research of recurrent mining of pattern to determine a minimal classical set of pattern having zero error algorithm known as MinRPset is implemented. MinRPset determines modest result which one can probably implement in experimental procedures for the mentioned situation and it takes ample time to conclude as soon as the count of periodic closed arrangement patterns is less than one million. MinRPset is highly utilizes memory space and time on numerous heavy datasets when the count of recurrent closed pattern is greater. To solve this difficulty, another algorithm called FlexRPset is utilized, that introduces one supplementary parameter K for facilitating users for making adjustment amongst efficiency and result size. An additional view to provide users to make intervene satisfaction is utilized. Some recurrent pattern mining usually develops a more count of periodic patterns that require a huge objective of understanding, visualizing and next investigation of the developed patterns. This increases the demand to find less count of repeatedly occurring patterns. In the proposed work, system classifies the patterns on using Machine Learning, which classifies the items browsed by user for finding exact pattern set. By using naive bayes technique, service provider can increase his sell and also the customers can be advised for exact items they need through exact pattern generation.
Authors and Affiliations
Preeti C. Tarange*
Keywords
Related Articles
- EP ID EP159530
- DOI 10.5281/zenodo.58653
- Views 72
- Downloads 0
A SURVEY ON DATA MINING TECHNIQUES, THEIR APPLICATION AND FUTURE SCOPE
Data mining techniques are rapidly growing and used by many researchers in past few years. The main advantage of using data mining is to extracts the knowledge from a huge amount of data which stores in variety of data r...
Using Letters Frequency Analysis in Caesar Cipher with Double Columnar Transposition Technique
In this paper we have some modification in Caesar Cipher Technique. We have proposed a method to enhancing the Caesar cipher for more efficient and secure. We use Relative Frequency of Letters in Alphabets. We arrange...
REVIEW ON GPS NAVIGATION SYSTEM FOR BLIND PEOPLE
For finding the way Blind People or Visually Impaired Person (VIP) mostly use white canes or Guide dogs for obstacle detection and avoidance, but these two traditional systems have some limitation. Hence it is need to...
PRODUCTION OF BIODIESEL FROM Calophyllum Inophyllum L. OIL BY LIPASE ENZYME AS BIOCATALYST
Production of biodiesel from Calophyllum inophyllum L oil through transesterification by enzyme as catalyst and methanol as solvent has been done. Determinations of molecular weight of lipase enzyme used SDS - P...
A Novel Method of Optimizing Lifetime and Reduction in Power Consumption in Wireless Sensor Network
Optimal power flow (OPF) problem has already been attempted as a static, optimization problem by adopting gradient based methods and more recently, no conventional ones, such as evolutionary algorithms. However,...