A Statistical Analysis of Big Web Market Data Structure Using a Big Dataset of Wines

The web market structure nowadays it's facing the issue of “big-data”. It is well known that big companies in the web like Amazon, Google, are trying to get information from their data. However similar interests have also other smaller companies. In this paper we considered a large dataset from a big company of wine, with white, rose and red wines (from Greece). Two regression techniques were considered the multivariate linear regression and logistic regression. The logistic regression gives stronger results in terms of the interpretation, outperforming the linear regression and we are expecting that others methods as neural networks and support vector machine improve the last one. We are dealing about this issue in another paper. Such models are useful to oenologist wine tasting evaluation and improve wine production. Furthermore, similar techniques can help in target marketing by modeling tastes from other markets.