Business intelligence can be said to be techniques and tools as acquisition, transforming raw data into meaningful and useful information for business analysis purposes. This study aims to build business intelligence in optimizing large-scale data based on e-metrics. E-metrics are data created from electronic-based customer behavior. As more and more large data sets become available, the challenge of analyzing data sets will get bigger and bigger. Therefore, business intelligence is currently facing new challenges, but also interesting opportunities, where can describe in real time the needs of the market share. Optimization is done using adaptive multivariate regression that can be address high-dimensional data and produce accurate predictions of response variables and produce continuous models in knots based on the smallest GCV value, where large and diverse data are simplified and then modeled based on the level of behavior similarity, basic measurements of distances, attributes, times, places, and transactions between social actors. Customer purchases will represent each preferred behaviour and a formula can be used to calculate the score for each customer using 7 input variables. Adaptive multivariate regression looks for customer behaviour so as to get the results of cutting the deviation which is the determining factor for performance on the data. The results show there are strategies and information needed for a sustainable business. Where merchants who sell fast food or food stalls are more in demand by customers.
Large-Scale Inference of Multivariate Regression for Heavy-Tailed and Asymmetric Data
