Online Product Recommendation Research Articles

Vulnerability of Gen-Z to E-Commerce Deception on Consumer's Belief Categories in Online Product Recommendations Systems

Purpose: Despite the convenience of the Online Product Recommendation (OPR) system, concerns persist regarding deception practices in online shopping, especially among consumers from the Generation Z demographic cohort, for which less attention has been given in the previous literature. Methodology/Approach: The impact of independent variables, namely Perceived Usefulness, Perceived Enjoyment, and Perceived Risk, on Susceptibility to e-commerce fraud among Gen Z consumers has been studied. By conducting a mall intercept survey in four major metropolitan cities in India with a usable sample of 488 responses, the study empirically tested the data using SmartPLS 4.0. Findings: The study concludes that Perceived Usefulness and Enjoyment positively influence continuous usage intention, while Perceived Deception and Perceived Risk are negatively connected with continuous usage intention. It depicts that the Gen Z consumer's belief categories formed during their early years as digital natives sharpen their alertness to deception practices. Research Limitation/Implication: The study may not be generalised to represent Gen Z consumers as it collected data only from those who visited malls in the metropolitan cities of India. Originality/Value of paper: This paper exclusively investigates the interrelation of belief categories to continuous usage intention. Further, the intersection of perceived Deception is proved in light of the E-commerce practices.

Analysis for Online Product Recommendation with recalling enhanced recurrent neural network-based sentiment

Analysis for Online Product Recommendation with recalling enhanced recurrent neural network-based sentiment

Order basket contents and consumer returns

AbstractAlthough lenient return policies can drive sales and customer loyalty, they have also resulted in enormous returns volumes and reverse logistics costs. Online retailers often feel compelled to offer free returns, but are then faced with numerous operational challenges, ranging from accurately forecasting returns volumes to identifying presales strategies to reduce the likelihood that a (costly) return occurs. In this research, we consider how the complementarity of the products within an order basket is related to consumer returns. By developing an understanding of the link between basket contents and returns, we can improve order‐level returns forecasts, while also providing insights into the effect of basket recommendations on the expected return rate. We take a multimethod approach to this problem. First, we use a stylized model to generate theoretical predictions regarding how within‐basket complementarity should influence return probability. Next, we propose a data‐driven measure of complementarity, degree of copurchase (DCP), which is based on the machine learning concept of association rule and is implementable using standard retail sales data. Finally, utilizing a unique data set provided by a leading online specialty retailer, we implement the DCP measure and test the predictions of our analytical model. We find, as expected, that there is a decreasing relationship between within‐basket complementarity and return probability. However, we also show that this decrease is convex, indicating that the return probability impact is more notable when the complementarity is increased from a lower base. Our results have practical implications for both reverse logistics planning and online product recommendations.

The role of recommendation sources and attribute framing in online product recommendations

As artificial intelligence (AI) is increasingly incorporated into online product recommendations (OPRs), investigating how an AI recommendation source influences consumer behavior has attracted widespread attention among scholars. Across five studies, this paper empirically examines the effect of the AI (vs. human) recommendation source on consumer responses from the perspective of vice and virtue frame products. The results show that when OPRs frame products as vice (vs. virtue), the AI (vs. human) recommendation source has negative effects on perceived warmth and competence (Study 2a) and eventually negatively influences purchase intention (Studies 1a and 2a), willingness to pay (Study 1b), and product attitude (Study 2b). However, humanized AI and AI-human hybrid improve the acceptance intention of vice frame product recommendations through different improvement paths (Study 3). This paper extends the research stream on the comparison of AI and humans and contributes to the literature on social perception, humanized intelligence, and augmented intelligence.

A novel framework for an intelligent deep learning based product recommendation system using sentiment analysis (SA)

Social media and e-commerce are the two most prominent and quickly expanding industries today. These two areas exhibit the greatest influence on platform users. Numerous new people sign up for these networks on a daily basis. This platform offers extremely quick user networking and communication. These platforms are used to create an online product-based recommender system that will help grow online business by recommending products. Online product recommendations are entirely dependent on the views, feedback, and comments of consumers. Online recommender systems have become a regular part of consumers’ everyday routines, with their widespread use observed in e-commerce, social networking platforms, and news websites. This paper offers a novel framework for product recommendation based on sentiment analysis (SA) and collaborative filtering (CF). The SA was performed using an LSTM-based model. On the basis of CF, two distinct recommendation systems were built. The proposed SA model was integrated with the best recommendation system to enhance the recommendations. The experimental findings showed that the proposed system for product recommendation outperformed the existing methods. The outcomes demonstrated the potential of combining CF and SA to improve consumer satisfaction and product recommendation in e-commerce systems.

Sentiment Analysis for Online Product Reviews and Recommendation Using Deep Learning Based Optimization Algorithm

Recently, online shopping is becoming a popular means for users to buy and consume with the advances in Internet technologies. Satisfaction of users could be efficiently improvised by carrying out a Sentiment Analysis (SA) of larger amount of user reviews on e-commerce platform. But still, it is a challenge to envision the precise sentiment polarity of the user reviews due to the modifications in sequence length, complicated logic, and textual order. In this study, we propose a Hybrid-Flash Butterfly Optimization with Deep Learning based Sentiment Analysis (HFBO-DLSA) for Online Product Reviews. The presented HFBO-DLSA technique mainly aims to determine the nature of sentiments based on online product reviews. For accomplishing this, the presented HFBO-DLSA technique applies data pre-processing at the preliminary stage to make it compatible. Besides, the HFBO-DLSA model uses deep belief network (DBN) model for classification. The HFBO algorithm is used as a hyperparameter tuning process to improve the SA performance of the DBN method. The experimental validation of the presented HFBO-DLSA method has been tested under a set of datasets. The experimental results reveal that the HFBO-DLSA approach surpasses recent techniques in terms of SA outcomes. Specifically, when compared to various existing models on the Canon dataset, the HFBO-DLSA technique achieves remarkable results with an accuracy of 97.66%, precision of 98.54%, recall of 94.64%, and an F-score of 96.43%. In comparative analysis, other approaches such as ACO, SVM, and NN exhibit poorer performance, while TextCNN, BiLSTM, and RCNN approaches yield slightly improved SA results.

Research on Predicting Customers' Next Purchase Based on Shopping Basket Data

Thanks to the rapid development of e-commerce and online shopping, a large amount of shopping basket data has been accumulated. How to mine the useful information in shopping cart data to predict customers' next purchase is an important problem in commercial data analysis, which is widely used in the fields of online advertising and product recommendation. In this paper, three prediction methods are proposed, including frequency-based prediction, rule-based prediction and similarity-based prediction. Moreover, evaluations and analysis of these three methods are conducted on the public dataset. It is found that the items that were frequently purchased by consumers in the past are more likely to continue to be purchased due of the higher prediction accuracy of the frequency-based methods. On the other hand, similarity-based item prediction methods also yielded good results because there is also a significant overlap in the items that similar users want to purchase. Therefore, it is concluded that in practical applications, frequency and similarity-based prediction methods can be applied to predict consumers' next purchases.

Graph Sample and Aggregate Attention Network optimized with Barnacles Mating Algorithm based Sentiment Analysis for Online Product Recommendation

Big data analytics is important in many businesses that use computing applications like real-time shopping and e-commerce. Big data is employed for promoting goods and improve connectivity betwixt retailers and customers. People today frequently utilize online promotions to learn about the best stores to visit in order to purchase higher-quality goods. This purchasing experience along opinion regarding the shopper’s shop can be observing through the customer-experience published across social media. A new customer requires information’s about manufacturing date (MRD), manufacturing price (MRP), offers, quality, and suggestions while searching a shop that are presented by prior customer experience. The product cover or label already has MRP and MRD. The product details have been predicted using a variety of methods, however the information is not correct. To overcome these issues, Graph Sample and Aggregate Attention Network (GSAAN) Optimized with Barnacles Mating Algorithm (BMA) based Sentiment Analysis (SA) is proposed for Online Product Recommendation (GSAAN-BMA-SA-OPR). The data’s are collected from the data set of Amazon Product recommendation. With the help of GSAAN with BMA classify product recommendation as excellent, good, very good, bad, very bad. The proposed model is implemented in MATLAB, its efficiency of the proposed system is examined under performance metrics, such as mean absolute error (MAE), mean squared error (MSE), mean absolute percentage error (MAPE), accuracy, F-Score, recall, precision. The proposed technique attains higher MAPE 23.77%, 65.3% and 10.42% and lower MSE 42.15%, 29.64% and 32.48% comparing to the existing models.

Self-attention generative adversarial capsule network optimized with atomic orbital search algorithm based sentiment analysis for online product recommendation

Self-Attention based Generative Adversarial Capsule Network optimized with Atomic orbital search algorithm based Sentiment Analysis is proposed in this manuscript for Online Product Recommendation (SFA-AGCN-AOSA-SA-OPR). Here, Collaborative filtering (CF) and product-product (P-P) similarity method is utilized for designing the new recommendation system. CF is employed for predicting the best shops and P-P similarity method is employed to predict the better product. Initially, the datas are gathered via Amazon Product recommendation dataset. After that, the datas are given to pre-processing. During pre-processing, Markov chain random field (MCRF) co-simulation is used to remove the unwanted content and filtering relevant text. The preprocessing output is fed to feature extraction. The features, like manufacturing date, Manufacturing price, discounts, offers, quality ratings, and suggestions or reviews are extracted using Gray level co-occurrence matrix (GLCM) window adaptive algorithm based feature extraction method. Finally, Self-Attention based Generative Adversarial Capsule Network (SFA-AGCN) categorizes the product recommendation as excellent, good, very good, bad, very bad. Atomic orbital search algorithm optimizes the SFA-AGCN weight parameters. The performance metrics, like accuracy, precision, sensitivity, recall, F-measure, mean absolute error (MAE), mean squared error (MSE), mean absolute percentage error (MAPE) is examined. The efficiency of the proposed method provides higher mean absolute percentage error 98.23%, 88.34%, 90.35% and 78.96% and lower Mean squared error 92.15%, 90.25%, 89.64% and 92.48% compared to the existing methods, such as sentiment analysis of online product reviews using DLMNN and future prediction of online product using IANFIS (DLMNN-IANFIS-SA-OPR), intelligent sentiment analysis approach using edge computing based deep learning technique (DCNN-SA-OPR), sentiment analysis for online product reviews in Chinese depending on sentiment lexicon and deep learning (CNN-BiGRU-SA-OPR) and sentiment analysis on product reviews depending on weighted word embedding and deep neural networks (CNN-LSTM-SA-OPR) respectively.

HOW DOES CUSTOMER RATING AFFECT PRODUCT PURCHASE?

A customers buying behavior is largely governed by the needs, preferences and tastes of the consumers for whom the product/service is ultimately purchased. Traditionally, consumers purchasing habits were solely relied on word of mouth. With the growth of the internet and technology, online consumer evaluations, product and service recommendations, and peer opinions play an increasingly important role in the customers decision-making process. According to the literature, buyers seek online product reviews for four reasons: information seeking, risk reduction, quality seeking, and social belonging. Before purchasing a product, the buyer actively gathers and integrates information from a variety of sources, the most cost-effective and immediately accessible of which are online review. These reviews will have both positive and negative effect on the consumers mindset. One bad review will drastically drop down the sales of the particular product as these opinions are considered as trustworthy and less risky than marketer information. This article will assist us in understanding how these ratings, reviews, and recommendations influence consumer product purchases and purchasing patterns. KEYWORDS: Purchasing pattern, Word of Mouth, Online reviews, Ratings, Sales

Reviewer reliability and XGboost whale optimized sentiment analysis for online product recommendation

Nowadays, people always use online promotions to know about best shops to buy the best products. This shopping experience and shopper’s opinion about the shop can be observed by the customer-experience shared on social media. A new customer when searching a shop needs information about manufacturing date (MRD) and manufacturing price (MRP), offers, quality, and suggestions which are only provided by the previous customer experience. Several approaches were used previously for predicting the product details, but no one approach provides accurate information. To overcome these issues, Reviewer Reliability and XGboost whale Optimized Sentiment Analysis for Online Product Recommendation is proposed in this manuscript.Initially, Amazon Product recommendation datathe data are preprocessed and given to XGboost Classifier that classifies the product recommendation result as, good, bad and average. Generally the XGboost Classifier does not reveal any adoption of optimization techniques for computing the optimal parameters for assuring accurate classification of product recommendation. Therefore in this work, proposed Whale optimization algorithm utilized to optimize the weight parameters of the XGboost. Then the proposed model is implemented in MATLAB. The proposed method attains 18.31%, 12.81%, 45.75%, 26.97% and 25.55% lower Mean Absolute error, 18.31%, 12.81%, 27.97%, 25.97%, and 25.55% higher Mean absolute percentage error and 15.31%, 10.33%, 25.86%, 22.86% and 15.22% lower Mean Square Error than the existing methods.

Enhanced Elman spike neural network based sentiment analysis of online product recommendation

The major task of Natural Language Processing (NLP) is Sentiments Analysis (SA) or Opinion Mining (OM). It captures the opinion, trust, and feelings of every user’s about the related products to find whether the attitude of users is positive, negative, or neutral. By this, companies can make necessary changes in the product for customer satisfaction. The majority of the existing approaches on sentiment analysis have inaccurate and takes more time. Therefore, in this manuscript, an Enhanced Elman Spike Neural Network based Sentiment Analysis of Online Product recommendation is proposed (EESNN-SA-OPR). Here, filtering Collaborative (FC) and product to product (P–P) similarity are used as new recommendation systems. The aim of Collaborative Filtering is “to predict the best shops and product–product​ similarity is to predict the best products”. Initially, the datas are taken from Amazon reviews database. Then the input data is pre-processed. The trilateral smoothing filtering (TRSF) is used to remove the content which is no more needed and texts related filtering. After that, the features like manufacturing price (MRP), Manufacturing date (MFD), discounts, offers, ratings in qualities are extracted by using Dominant Gradient Local Ternary Pattern descriptor (DGLTPD) technique. At last, enhanced Elman spike neural network classifies the product recommendation as excellent, good, very good, bad and very bad. The proposed method is executed in MATLAB and its performance is analyzed under some performance metrics, like mean absolute error (MAE), mean squared error (MSE), mean absolute percentage error (MAPE), accuracy, F-Score, recall and precision. The proposed EESNN-SA-OPR method provides 23.14%, 15.96%, 31.54% higher accuracy and 12.33%, 21.31%, 41.09% lower mean absolute error compared with the existing techniques, like Sentiment analysis of online product reviews utilizing DLMNN and future prediction of online product utilizing IANFIS (DLMNN-SA-OPR), a machine learning-based sentiment analysis of online product reviews with new term weighting along feature selection method (LSIBA-ENN-SA-OPR) and Sentiment Analysis on the Reviews of Online Product utilizing Optimized RNN-LSTM along Support Vector Machine (RNN-LSTM-SA-OPR) respectively.

Customer's decision and affective assessment of online product recommendation: A recommendation-product congruity proposition.

Online product recommendation (OPR) systems have gained prominence in the context of e-commerce over the past years. Despite the increased research on OPR use, less attention has been paid to examining how decision and affective assessment of the OPR are contingent upon the product type. This study proposes and examines a recommendation-product congruity proposition based on cognitive fit and schema congruity theories. The proposition states that when the content (i.e., a stimulus-based schema) of the OPR [either system-generated recommendation (SGR) or a consumer-generated recommendation (CGR)] matches the brain-stored schema initiated by a particular product (either a search product or an experienced product), then a consumer would use a schema-based information assessment strategy and experience favorable decision and affective assessment of the OPR. This then affects consumers’ intentions to purchase and reuse OPR. The proposition is tested via a 2 × 2 between-respondents factorial design of a cross-sectional survey with 482 Amazon customers. The results support the following two matching conditions of the proposition: (1) SGR describing a search product and (2) CGR explaining an experienced product, which might lead customers to perceive lower decision effort, greater decision quality, and higher enjoyment with the OPR that subsequently have a significant impact on their intentions to purchase and reuse OPR. This study expands our understanding of how recommendation-product congruence influences the consumer’s decision and affective assessment behavior and provides practical implications for the identification and presentation of the recommendation type and product type for a better customer decision.

Product Brokering Efficiency as a mediator of Online Product Recommendation and Customer Loyalty

Marketing strategies in e-commerce have a main goal, that is to pursue customer loyalty. Sociolla is an e-commerce company that sells cosmetic products and has a recommendation feature to make it easier for customers during the shopping process. These recommendations can trigger customer satisfaction and generate loyalty. The purpose of this study was to examine the correlation between online product recommendation and customer loyalty with product brokering efficiency as a mediator. 179 Sociolla customers were recruited in this study using convenience sampling. The data were analyzed using the SPSS-Process Hayes model 4. Results showed that perceived decision quality acts as a mediator in the relationship between enablers and customer loyalty (β = .20, [ .13; .27]). It can be concluded that recommendations that are comprehensive, clear, and meet the customer needs will make it easier for customers to make purchasing decisions, which ultimately leads the customers to form loyalty toward the products.

Online product recommendation system using gated recurrent unit with Broyden Fletcher Goldfarb Shanno algorithm

In recent decades, online product recommendation system has become a main channel for enterprise promotion, because it is rapidly used across several aspects of ecommerce and online media. However, dealing with the customer feedbacks in text format as an unstructured data is a challenging task, because it is hard to analyze and interpret the information. In this research work, matrix factorization and non-negative matrix factorization methods are applied in gated recurrent unit to predict the long and global time interested products of the users. The factorization methods generate the latent user and product features in gated recurrent unit for underlying the interaction between users and products. Additionally, the obtained latent user and product features are feed to Broyden Fletcher Goldfarb Shanno algorithm to recommend the final product to the customers. In this paper, gated recurrent unit gives differentiable dependencies on predicted state. To resolve the problems of non-linear constrains in gated recurrent unit, the Broyden Fletcher Goldfarb Shanno algorithm is applied in this work. Simulation results showed that the proposed algorithm achieved better performance in product recommendation compared to collaborative filtering, fuzzy c means, k-means clustering and quantum inspired possibilistic fuzzy C-means on amazon customer review database in terms of precision, recall, and accuracy.

Power‐average‐operator‐based hybrid multiattribute online product recommendation model for consumer decision‐making

This study develops a power-average-operator-based hybrid multiattribute online product recommendation model that considers the consumer's risk attitude to rank categoric product options as a complement to existing recommender systems. Online production recommendation plays a key role in the development of e-commerce, and can greatly improve consumers' shopping experiences. However, few online shopping sites provide interactive decision aids for consumers such that they can articulate their preferences towards multiple selection attributes with the purpose of mitigating choice difficulty and improving decision quality. Additionally, consumers' risk attitudes to online shopping dramatically impact their product choices. In the model proposed in this paper, the risk attitude-based power average (RAPA) operator is used to integrate the risk attitude of the decision-maker into the information fusion process of multiple attribute decision-making. Subsequently, the risk attitude function, with several basic types, is introduced to quantify the risk attitude of the decision-maker for use in the RAPA operator. A proportional hesitant fuzzy 2-tuple linguistic term set (PHF2TLTS) is constructed by incorporating a binary of linguistic information aiming to comprehensively analyze the hybrid product information. With a focus on the information fusion process, the proportional hesitant 2-tuple linguistic RAPA operator and weighted proportional hesitant 2-tuple linguistic RAPA operator are introduced to aggregate a given set of PHF2TLTSs. The validity of the proposed model is demonstrated using an illustrative example, a comparison with existing approaches and detailed explanations of the performance differences.

Incorporating Similarity Measures to Optimize Graph Convolutional Neural Networks for Product Recommendation

With the ever-growing amount of online data and information, recommender systems are becoming overwhelmingly popular as an adequate approach for overcoming the challenge of information overload. Artificial Intelligence (AI) and Deep Learning (DL) have accumulated significant interest in many research areas, and recommender systems are one of them. In this paper, a Graph Convolutional Neural Network (GCNN)-based approach was used for online product recommendation. Graph-based methods have undergone substantial consideration for several recommendation tasks, with effective results. However, handling the computational complexities and training large datasets remain a challenge for such a model. Even though they are useful, the excessive measure of the model’s boundaries obstructs their applications in real-world recommender frameworks to a great extent. The recursive way of generating neighbor node embeddings for each node in the graph makes it more challenging to train a deep and large GCNN model. Therefore, we propose a model that incorporates measures of similarity between two different nodes, and these similarity measures help us to sample the neighbors beforehand. We estimate the similarity based on their interaction probability distribution with other nodes. We use KL divergence on different probability distributions to find the distance between them. This way, we set a threshold criterion for neighbor selection and generate other clusters. These clusters are then converted to subgraphs and are used as input for the proposed GCNN model. This approach simplifies the task of neighbor sampling for GCNN, and hence, we can observe a significant improvement in the computational complexity of the GCNN model. Finally, we compared the results with those for the previously proposed OpGCN model, basic GCNN model, and other traditional approaches such as collaborative filtering and probabilistic matrix factorization. The experiments showed that the complexity and computational time were decreased by estimating the similarity among nodes and sampling the nodes before training.

The Effect Of Online Product Recommendation System On Consumer Behavior: Vietnameses E-Commerce Websites.

Online Product Recommendation (OPR) systems offered by Ecommerce website are crucial in providing real-time product recommendations based on customers’ past behavior and references, contributing a significant impact on increasing purchasers’ satisfaction, boosting websites’ revenue through up-selling or cross-selling, and settling a long-term relationship between websites and customers. This research model was developed to investigate how positive and negative characteristics of OPR’s quality influence purchasers’ decision-making process and their time spent while surfing E-commerce websites, and how these two factors conclusively affect customer loyalty. The results point out that customer loyalty is significantly influenced by decision making quality and is negatively affected by customer time wasted. There are positive relationships between Content relevance, vividness and the DMQ; positive association between deceptiveness, information overload and customer time wasted. The results of this research can be utilized by E-commerce websites to optimize their Product Recommendation system to gain more customer loyalty and be a strong player in the market

Product Bundle Recommendation and Pricing: How to Make It Work?

Product bundling is a common marketing strategy for cross-selling in multiproduct firms. Motivated by settings in online product recommendation, we propose a new approach, dubbed recommendation and pricing (BRP), to enhance the performance of recommendation system. BRP keeps all the separately priced products in the recommended set, and adds a subset of products as a new with a discounted price to customers. This approach extends pure bundling (PB), where all the products are sold in a single with a discounted price to customers. Although PB can be more profitable than component pricing (CP) where products are priced and sold separately, it can be inferior to in the presence of high marginal cost. We show that such a simple CP + one bundle scheme can be more profitable than both PB and CP, and is near optimal in many environments. BRP improves by extracting the deadweight loss, but retains the profitability of when some products have relatively high marginal costs. However, finding the optimal BRP solution is often intractable. We develop a new approximation to this problem and use a Bayesian optimization algorithm to optimize the selection and pricing decisions. Extensive numerical results show that our algorithm outperforms other common heuristics. More importantly, by simply adding one more option to the common mechanism, our results show that BRP tends to significantly increase both the monopolist's profit and customers' utility as compared with and PB.

Modeling low- and high-order feature interactions with FM and self-attention network

Click-Through Rate (CTR) prediction has always been a very popular topic. In many online applications, such as online advertising and product recommendation, a small increase in CTR will bring great returns. However, CTR prediction has always faced several challenges. A large number of users and items and the different sizes of the feature space of different data types lead to high-dimensional and sparse input, and high-order feature interactions rely too much on expert knowledge and are very time-consuming. In this paper, we build a novel model called multi-order interactive features aware factorization machine (MoFM) for CTR prediction. To effectively capturing both low-order and high-order interactive features, three different types of prediction models are integrated, of which logistic regression (LR) and factorization machine (FM) model the original features and 2-order interactive features respectively, and a multi-head self-attention network with residual connections is used to automatically identify high-value high-order feature combinations. There is also an embedding layer in the model to realize a unified embedding processing of different data types, avoiding diversification, sparsity, and high dimensionality of features. Since, feature engineering is not required, we can carry out end-to-end model learning. Experiments on three public datasets show the superiority of the proposed model over the state-of-the-art models, and the flexibility and scalability of the model structure have also been verified.