01-06-2013, 01:06 PM
Informative Knowledge Discovery using Combined Mining
Informative Knowledge.ppt (Size: 408.5 KB / Downloads: 26)
INTRODUCTION
Enterprise applications consist of large volumes of data, mining
such data causes space complexity and time complexity.
Traditional data mining techniques are available to identify
the homogeneous features of patterns in the data sources.
Combined Mining approach apply Location based search alignment
algorithm (LBSA) and Cluster kinship search technique (CKST)
to generate the incremental cluster patterns.
Combined mining is applied on synthetic mobile transaction dataset.
Clustering is one of the most widely used technique
in data mining and knowledge discovery and has tremendous applications like business, science and other domains.
In the global mobile computing environments,
mobile users may request diverse kinds of services and applications from arbitrary locations at any time via on networks.
Discovery of mobile user behavior becomes complex
in large transaction databases.
Dataset Extraction
Get informative patterns from data source.
A temporal dataset consists of multiple transactions of users.
To retrieve the similar patterns LBST used.
Generating Similar Patterns
Similar patterns get over by applying LBST on mobile transaction data set.
Different users who are accessing the services at different times from same location are created as one pattern and stored in similarity matrix.
Building Similarity Matrix
In frequent pattern mining, the transactions containing same pattern would look similar in appearance to each other.
we proposed a similarity matrix based method to mining maximal frequent patterns from large database.
Find Similarity of Patterns based on Cluster Analysis
In similar pattern mining, the user patterns contain some temporal information generated at different timings.
Users accessing service at same time from particular location have same similarity value calculated from similarity matrix.
Find maximum similarity value of all user patterns.
Conclusions
Combined mining is better than individual data mining methods to
retrieve informative patterns from dataset.
A tool Informative Knowledge Discovery using combined mining to
get incremental cluster patterns as informative patterns. Using these
patterns Prediction of user behavior and interestingness calculated.