Bacterial colony algorithm with adaptive attribute learning strategy for feature selection in classification of customers for personalized recommendation

摘要: Abstract This paper investigates a new bacterial colony-based feature selection algorithm to improve the classification accuracy of customers for personalized products recommendation. An attribute learning strategy is developed in this study update related population. Specifically, features can be weighted according their historic contributions both individual- and group-based subsets. Additionally, frequency appearance also recorded candidates diversity distribution avoid over-fitting. Based on weight-based indexes records, performance subsets are enhanced by replacing being repeatedly appeared same vector. To explore feasibility proposed method missing problems, objective optimization minimize error using acceptable number features. K-Nearest Neighbor employed as technique cooperate with method. The effectiveness demonstrated performing test datasets from UCI machine repository real-world data Amazon customer reviews products. Compared other seven selections methods, outperforms algorithms achieving higher rate smaller