Classification and optimization of product review information using soft computing models

A soft computing framework to classify and optimize text-based information extracted from customers' product reviews is proposed in this paper. The soft computing framework performs classification and optimization in two stages. Given a set of keywords extracted from unstructured text-based product reviews, a Support Vector Machine (SVM) is used to classify the reviews into two categories (positive and negative reviews) in the first stage. An ensemble of evolutionary algorithms is deployed to perform optimization in the second stage. Specifically, the Modified micro Genetic Algorithm (MmGA) optimizer is applied to maximize classification accuracy and minimize the number of keywords used in classification. Two Amazon product reviews databases are employed to evaluate the effectiveness of the SVM classifier and the ensemble of MmGA optimizers in classification and optimization of product related keywords. The results are analyzed and compared with those published in the literature. The outputs potentially serve as a list of impression words that contains useful information from the customers' viewpoints. These impression words can be further leveraged for product design and improvement activities in accordance with the Kansei engineering methodology.

Cumulative restricted Boltzmann machines for ordinal matrix data analysis

Ordinal data is omnipresent in almost all multiuser-generated feedback - questionnaires, preferences etc. This paper investigates modelling of ordinal data with Gaussian restricted Boltzmann machines (RBMs). In particular, we present the model architecture, learning and inference procedures for both vector-variate and matrix-variate ordinal data. We show that our model is able to capture latent opinion profile of citizens around the world, and is competitive against state-of-art collaborative filtering techniques on large-scale public datasets. The model thus has the potential to extend application of RBMs to diverse domains such as recommendation systems, product reviews and expert assessments.

Crowdsourcing customer needs for product design using text analytics

In product design and engineering, identifying customer needs is the foundation for designing and producing a successful product. Traditionally, a range of techniques have been employed to elicit customer needs. A relatively new technique for identifying customer needs is ‘crowdsourcing’. An emerging area of research is the crowdsourcing of customer needs from online product review sites. This paper proposes a simple process for crowdsourcing customer needs for product design using text analytics. The analysis/visualization method is presented in detail. The text content of online customer reviews for a popular product is collected and processed using text analytics software. A published case study identifying expressed customer needs for the same generic product, collected via conventional means, is used to successfully validate the findings from the text analytics method.