Iterated Data Mining Techniques on Embedded Vector Modeling
Classification and prediction problems of transaction databases are well known data-mining problems. Their importance becomes further noticeable at the time of information data explosion. The existence of high volume transactional data provides us the challenge of exploring the valuable and meaningful hidden information. The classic approaches, such as association analysis and cluster analysis, always request human interventions either in predefining the logical correlations or in setting the weights for attribute parameters ([HK]). Also in the most cases the binary results are presented with prematurely chosen thresholds.
KeywordsMatch Rate User Space Action Item Transaction Database Space Embedding
Unable to display preview. Download preview PDF.