Contextual Analysis of Transactional Data

摘要

Data analysis is often attempted by business stakeholder on transactional databases, that are maintained within the overall enterprise system, in order to understand patterns of user behavior, transaction types, and their impact on the utility and value of interest to the enterprise. Statistical and machine learning techniques are often used to make predictions and decisions. In order to guide the business stakeholder ask the right questions and guide the query processor in fetching the most accurate answers to the questions we introduce formal approaches in which the five dimensional context information What?, Why?, Where?, When?, Who? are captured. When the transactions are contextually qualified both querying and analysis of data on their responses will lead to efficient and accurate discovery of interesting patterns for decision making.