Data Mining Sales Optimizations Using Sequential Minimal Optimization Algorithm

Dedy Kurniadi, Sam Farisa Chaerul Haviana

Abstract


Tightness of business nowadays requires businessman to be able to develop their business to compete with the other companies, this study was conducted to obtain data accurate on the type of clothing combinations that are favored by the consumers to optimize sales at convection companies, using data mining methods and technique of classification this data is classify into four classes namely, well-liked, liked, enough and dislike. To solve classification problems, this study used Sequential Minimal Optimization (SMO), SMO Algorithm can solve quadratic programming problems without requiring a large matrix and to solving the optimization SMO selected from the smallest optimization in every steps. Optimum accuracy obtained in this study were obtained from Correctly Classified Instance of 80.9% from 3072 record set of well-liked classes that is the class with type of combinations clothes polo and embroidery, then the level of measurement of consistency coefficient values using kappa statistic obtained for 0.73% where the data in the class showed a consistent value, from these data type are most well-liked combinations can optimize sales by 70.3%


Refbacks

  • There are currently no refbacks.


View JTI Stats