An Efficient Data Structure for Data Mining Techniques

Abstract

Today s world deals with several petabytes of data and the utilization of data increases every single data. Now-a-days, data collection from the users has become common to improve the functionality or the Quality of Service (QoS) of the application. In most of the cases, the business communities collect the customer information and get to know the personal interests. This is to improve the business experience and the user satisfaction. newlineSimilarly, the medical world records the health data for future references. The health data can be the medical reports such as scan reports or other readings such as Blood Pressure (BP), Haemoglobin level and so on. All the records associated with the patient are stored for ensuring easy analysis and thereby better diagnosis. newlineHence almost in all the domains, data plays a vital role and now-a-days, e-data has become a significant requirement of many applications. All the aforementioned data are stored digitally and the analysis is performed to improvise the business experience or the final outcome of the process. Though the collected data are stored in the database, it is extremely difficult to analyse the vast amount of data immediately. newlineThe data analysis can easily be performed, when the data is organized in a better way. Understanding the importance of data organization, this research work presents three different solutions dealing with different problems, such as frequent data mining, Multi-Label Classification and disease prediction. Each considered problem is different from the other and all the proposed solutions utilize a common data structure for organization. newlineInitially, this research work presents a frequent pattern mining approach based on One Time Scan plus Pyramid. This work is based on a single round of scanning and the data organization is carried out with the help of pyramid data structure. The Pyramid Data Structure holds the data with respect to the degree of combination and the occurrence frequency of the corresponding combination.