The Recommending Agricultural Product Sales Promotion Mode in E-Commerce Using Reinforcement Learning with Contextual Multiarmed Bandit Algorithms

Jyh-Yih Hsu,Jia-You Hsieh,Wei-Kuo Tseng,Chao-Jen Chang,Huan Chen,Jason C Hung

doi:10.1155/2020/8836000

Jyh-Yih Hsu, Jia-You Hsieh + Show 4 more

Open Access

https://doi.org/10.1155/2020/8836000

Copy DOI

Abstract

In recent years, sales of agricultural products in Taiwan have been transformed into electronic marketing, and agricultural products with better consumer orientation have been recommended, and farmers’ income has been improved through sales websites. In the past, A/B testing was used to determine the degree of preference for website solutions, which required a large number of tests for evaluation, and could not respond to environmental variables that made it difficult to predict the actual recommendation in advance. Therefore, in this study, the reinforcement learning model combined with different contextual Multiarmed Bandit algorithms can be tested in data sets of different complexity, which can actually perform well in changing products. It is helpful to predict the preferences of the promotion model.

Highlights

Governments of various countries have spared no effort to promote electronic sales of agricultural products. ey have cooperated with private businesses to set up websites for selling agricultural products [1] and guided local farmers’ associations to establish online shopping malls [2]. is shows that it is important for the electronic sales of agricultural products. e same agricultural products sales websites are facing the problem of how to market and promote agricultural products websites
We can know that the α constant in LinUCB, Hybrid-LinUCB, CoLin, and hLinUCB is related to the characteristics of the data set. e constant α does have to be larger or smaller but does have setting which was based on the current environment. erefore, it must be noted that, in each performance of the algorithm, the current sample will determine the difference in user characteristics, and there will be a certain degree of uncertainty in the simulation test
We can know that the LinUCB algorithm is highly recognizable for the linear relationship between user characteristics and product characteristics and can be used in most cases. e Hybrid-LinUCB algorithm has common environmental characteristics, and it has better results for changing products and can avoid the problem of cold start of expected value

Summary

Introduction

Governments of various countries have spared no effort to promote electronic sales of agricultural products. ey have cooperated with private businesses to set up websites for selling agricultural products [1] and guided local farmers’ associations to establish online shopping malls [2]. is shows that it is important for the electronic sales of agricultural products. e same agricultural products sales websites are facing the problem of how to market and promote agricultural products websites.erefore, in e-commerce websites [3], to increase sales has always been an important issue for website operation. Governments of various countries have spared no effort to promote electronic sales of agricultural products. Is shows that it is important for the electronic sales of agricultural products. In the field of e-commerce, understanding consumer characteristics and behaviors and to sell recommending products is one of the important goals in e-commerce websites. In order to predict the will of the consumer, we must rely on the model and find the relevance from different features, such as age, gender, and region. Because of the user characteristics, we cannot see the correlation with consumer promotion preferences manually, so we must rely on models to find the correlation from the features. Reinforcement learning is the situational dobby algorithm that will be used in this article. Reasons for choosing reinforcement learning in this article are as follows:

Objectives

Methods

Results

Conclusion