Developing Data-Conscious Deep Learning Models for Product Classification

Yijin Kim,Junho Shim,Hong Joo Lee

doi:10.3390/app11125694

Abstract

In online commerce systems that trade in many products, it is important to classify the products accurately according to the product description. As may be expected, the recent advances in deep learning technologies have been applied to automatic product classification. The efficiency of a deep learning model depends on the training data and the appropriateness of the learning model for the data domain. This is also applicable to deep learning models for automatic product classification. In this study, we propose deep learning models that are conscious of input data comprising text-based product information. Our approaches exploit two well-known deep learning models and integrate them with the processes of input data selection, transformation, and filtering. We demonstrate the practicality of these models through experiments using actual product information data. The experimental results show that the models that systematically consider the input data may differ in accuracy by approximately 30% from those that do not. This study indicates that input data should be sufficiently considered in the development of deep learning models for product classification.

Highlights

Internet commerce has become more active as internet distribution networks have grown
Based on the data-conscious method proposed in this paper, we compared and evaluated the performance of each model trained by plugging in the convolutional neural network (CNN) and bidirectional long short-term memory (Bi-LSTM) models presented in the previous section
These results demonstrate that if the CNN deep learning model is used indiscriminately without considering the data as a product classification model, its accuracy can differ by approximately 67.74% from that of the CNN model that effectively considers the data in both selection and transformation phases

Summary

Introduction

Internet commerce has become more active as internet distribution networks have grown. The appropriate product category can be identified if text data extracted from past data of used-product transactions are refined and used to train a deep learning model. This is a technical solution to the problem of inaccurate classification. This study found that deep learning model development and hyperparameter tuning should be integrated for the resulting model to be conscious of the product information data [5]. We adopted a methodology that performs the required partial transformation and adds layers to fit the data in these typical deep learning models These models are established to be suitable for the product classification of product information text data.

Related Work

Case Study of Filtering Efficiency

Findings

Conclusions