No abstract

KANO model classification is helpful for us to recognize customer needs and to improve their satisfaction. The traditional method uses standard questionnaires to conduct surveys, classifies product attributes according to the survey results. However with the increase of product complexity and the speed of product iteration, the method of survey is more and more unable to meet our analysis needs; coupled with the increasing number of customers who do not want to give feedback for questionnaires, low responds ratio rate leads poor feedback quality which affects the reliability of the research results. Although many studies are about KANO model classification, few of them focus on how to improve responds ratio rate. This article creates a new method for KANO model classification. By collecting customer reviews and rating score, we build up regression model between the score and the degree to which product attributes meet user needs according to their text expression. Based on the curve shape of the model coefficients and the value of the coefficient we can identify which KANO classification will a product attribute belongs to. The experiment study for gaming notebook has proved that this method is efficient and can be widely used in other products. We call this method as KKMA (Kano, K-means, MDS, Ad boost).

The lack of high-quality labeled training data has been one of the critical challenges facing many industrial machine learning tasks. To tackle this challenge, in this paper, we propose a semi-supervised learning method to utilize unlabeled data and user feedback signals to improve the performance of ML models. The method employs a primary model M ain and an auxiliary evaluation model Eval, where M ain and Eval models are trained iteratively by automatically generating labeled data from unlabeled data and/or users feedback signals. The proposed approach is applied to different text classification tasks. We report results on both the publicly available Yahoo! Answers dataset and our e-commerce product classification dataset. The experimental results show that the proposed method reduces the classification error rate by 4% and up to 15% across various experimental setups and datasets. A detailed comparison with other semi-supervised learning approaches is also presented later in the paper. The results from various text classification tasks demonstrate that our method outperforms those developed in previous related studies.