• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.01a
• /
• pp.31-33
• /
• 2022

쇼핑몰 리뷰 생성기는 사용자로 하여금 사용자를 대신해서 리뷰를 생성할 수 있는 기술이고, 옷 상태, 배송 상태, 사이즈와 관련된 세 가지의 카테고리를 이용하여 부분마다 점수를 부여하여 점수에 맞는 리뷰를 생성할 수 있도록 하는 기술이다. 해당 리뷰 생성기는 점수마다 생성되는 리뷰가 달라지기 때문에 다양한 리뷰 생성을 원하는 웹, 앱 쇼핑몰 사이트에서 적용이 가능한 기술이다. 본 논문에서는 KoGPT2를 이용한 리뷰 생성과 카테고리와 점수에 따른 다르게 생성되는 리뷰의 방식을 제안한다. 그리고 두 방식을 결합한 리뷰 생성의 방식을 제안한다. 제안하는 방식들은 카테고리고리 마다 학습하는 모델을 다르게 적용하고 있다.

• The Journal of the Korea Contents Association
• /
• v.20 no.11
• /
• pp.228-236
• /
• 2020

As of March 2020, this study divided the applications for young children installed on Android-based smart-phones in Korea into top and bottom groups according to review scores, and selected 30 applications each, conducted content analysis and application evaluation, and looked at differences between groups. Through this, by providing objective information on the smart-phone application for young children, it is intended to help parents and early childhood education professionals select high-quality applications, and to present ideas and directions for developing applications suitable for development to application developers. As a result of application content analysis, only data presentation type, simulation type, and game type were found in all the top and bottom groups as for the application type. There was a difference in order. In the case of app purchase cost, the top group in the review score was evenly distributed from the low price to the high price of 100,000 won or more, while the bottom group had few high-priced applications. On the other hand, as a result of application evaluation, a significant difference was found in the entire evaluation score, including all functional elements and all content elements, between the top and bottom groups of the review score. In the case of detailed sub-factors, significant differences were shown in all factors except 'technicality' of functional elements.

• Journal of Intelligence and Information Systems
• /
• v.22 no.3
• /
• pp.71-89
• /
• 2016

Sentiment analysis is used for identifying emotions or sentiments embedded in the user generated data such as customer reviews from blogs, social network services, and so on. Various research fields such as computer science and business management can take advantage of this feature to analyze customer-generated opinions. In previous studies, the star rating of a review is regarded as the same as sentiment embedded in the text. However, it does not always correspond to the sentiment polarity. Due to this supposition, previous studies have some limitations in their accuracy. To solve this issue, the present study uses a supervised sentiment classification model to measure a more accurate sentiment polarity. This study aims to propose an advanced sentiment classifier and to discover the correlation between movie reviews and box-office success. The advanced sentiment classifier is based on two supervised machine learning techniques, the Support Vector Machines (SVM) and Feedforward Neural Network (FNN). The sentiment scores of the movie reviews are measured by the sentiment classifier and are analyzed by statistical correlations between movie reviews and box-office success. Movie reviews are collected along with a star-rate. The dataset used in this study consists of 1,258,538 reviews from 175 films gathered from Naver Movie website (movie.naver.com). The results show that the proposed sentiment classifier outperforms Naive Bayes (NB) classifier as its accuracy is about 6% higher than NB. Furthermore, the results indicate that there are positive correlations between the star-rate and the number of audiences, which can be regarded as the box-office success of a movie. The study also shows that there is the mild, positive correlation between the sentiment scores estimated by the classifier and the number of audiences. To verify the applicability of the sentiment scores, an independent sample t-test was conducted. For this, the movies were divided into two groups using the average of sentiment scores. The two groups are significantly different in terms of the star-rated scores.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.11 no.2
• /
• pp.181-188
• /
• 2018

In this paper, we propose an analysis scheme of customer spending pattern using text mining. In proposed consumption pattern analysis scheme, first we analyze user's rating similarity using Pearson correlation, second we analyze user's review similarity using TF-IDF cosine similarity, third we analyze the consistency of the rating and review using Sendiwordnet. And we select the nearest neighbors using rating similarity and review similarity, and provide the recommended list that is proper with consumption pattern. The precision of recommended list are 0.79 for the Pearson correlation, 0.73 for the TF-IDF, and 0.82 for the proposed consumption pattern. That is, the proposed consumption pattern analysis scheme can more accurately analyze consumption pattern because it uses both quantitative rating and qualitative reviews of consumers.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.10
• /
• pp.87-94
• /
• 2020

Social reviews such as SNS feeds and blog articles have been widely used to extract keywords reflecting opinions and complaints from users' perspective, and often include proper nouns or new words reflecting recent trends. In general, these words are not included in a dictionary, so conventional morphological analyzers may not detect and extract those words from the reviews properly. In addition, due to their high processing time, it is inadequate to provide analysis results in a timely manner. This paper presents a method for efficient keyword extraction from social reviews based on the notion of cohesion scoring. Cohesion scores can be calculated based on word frequencies, so keyword extraction can be performed without a dictionary when using it. On the other hand, their accuracy can be degraded when input data with poor spacing is given. Regarding this, an algorithm is presented which improves the existing cohesion scoring mechanism using the structure of a word tree. Our experiment results show that it took only 0.008 seconds to extract keywords from 1,000 reviews in the proposed method while resulting in 15.5% error ratio which is better than the existing morphological analyzers.

• The Journal of the Korea Contents Association
• /
• v.22 no.1
• /
• pp.58-73
• /
• 2022

The purpose of this study is to construct a predictable estimation model that reflects the relationship between the variables of mobile app performance and to verify how app reviews affect app performance. In study 1 and 2, the relationship between app performance indicators was derived using correlation analysis and random forest regression estimation of machine learning, and app performance estimation modeling was performed. In study 3, sentiment scores for app reviews were by using sentiment analysis of text mining, and it was found that app review sentiment scores have an effect one lag ahead of the number of daily installations of apps when using multivariate time series analysis. By analyzing the dissatisfaction and needs raised by app performance indicators and reviews of apps, companies can improve their apps in a timely manner and derive the timing and direction of marketing promotions.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.12
• /
• pp.145-153
• /
• 2020

Recently, keyword extraction from social big data has been widely used for the purpose of extracting opinions or complaints from the user's perspective. Regarding this, our previous work suggested a method to improve accuracy of keyword extraction based on the notion of cohesion scoring, but its accuracy can be degraded when the number of input reviews is relatively small. This paper presents a method to resolve this issue by applying simplified morphological analysis as a postprocessing step to extracted keywords generated from the algorithm discussed in the previous work. The proposed method enables to add analysis rules necessary to process input data incrementally whenever new data arrives, which leads to reduction of a dictionary size and improvement of analysis efficiency. In addition, an interactive rule adder is provided to minimize efforts to add new rules. To verify performance of the proposed method, experiments were conducted based on real social reviews collected from online, where the results showed that error ratio was reduced from 10% to 1% by applying our method and it took 450 milliseconds to process 5,000 reviews, which means that keyword extraction can be performed in a timely manner in the proposed method.

• The Journal of Society for e-Business Studies
• /
• v.16 no.3
• /
• pp.93-112
• /
• 2011

For the success of an online retail market, it is important to allow consumers to get more helpful reviews by figuring out the factors determining the helpfulness of online reviews. On the basis of elaboration likelihood model, this study analyzes which factors determine the helpfulness of reviews and how the factors affecting the helpfulness of an online consumer review differ for product price. For this study, 75,226 online consumer reviews were collected from Amazon.com. Furthermore, additional information on review messages was also gathered by carrying out a content analysis on the review messages. This study shows that both of peripheral cues such as review rating and reviewer's credibility and central cues such as word count of review message and the proportion of negative words influence the helpfulness of review. In addition, the result of this study reveals that each consumer focuses on different information sources of reviews depending on the product price.

• Journal of Internet Computing and Services
• /
• v.25 no.4
• /
• pp.47-56
• /
• 2024

This paper proposes a method for fine-tuning BERT-based natural language processing models to perform sentiment analysis on Korean review data. By varying the input sequence length during this process and comparing the performance, we aim to explore the optimal performance according to the input sequence length. For this purpose, text review data collected from the clothing shopping platform M was utilized. Through web scraping, review data was collected. During the data preprocessing stage, positive and negative satisfaction scores were recalibrated to improve the accuracy of the analysis. Specifically, the GPT-4 API was used to reset the labels to reflect the actual sentiment of the review texts, and data imbalance issues were addressed by adjusting the data to 6:4 ratio. The reviews on the clothing shopping platform averaged about 12 tokens in length, and to provide the optimal model suitable for this, five BERT-based pre-trained models were used in the modeling stage, focusing on input sequence length and memory usage for performance comparison. The experimental results indicated that an input sequence length of 64 generally exhibited the most appropriate performance and memory usage. In particular, the KcELECTRA model showed optimal performance and memory usage at an input sequence length of 64, achieving higher than 92% accuracy and reliability in sentiment analysis of Korean review data. Furthermore, by utilizing BERTopic, we provide a Korean review sentiment analysis process that classifies new incoming review data by category and extracts sentiment scores for each category using the final constructed model.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.41-63
• /
• 2023

As the online e-commerce market growing, the need for a recommender system that can provide suitable products or services to customer is emerging. Recently, many studies using the sentiment score of online review have been proposed to improve the limitations of study on recommender systems that utilize only quantitative information. However, this methodology has limitation in extracting specific preference information related to customer within online reviews, making it difficult to improve recommendation performance. To address the limitation of previous studies, this study proposes a novel recommendation methodology that applies deep learning technique and uses various linguistic factors within online reviews to elaborately learn customer preferences. First, the interaction was learned nonlinearly using deep learning technique for the purpose to extract complex interactions between customer and product. And to effectively utilize online review, cognitive contents, affective contents, and linguistic style matching that have an important influence on customer's purchasing decisions among linguistic factors were used. To verify the proposed methodology, an experiment was conducted using online review data in Amazon.com, and the experimental results confirmed the superiority of the proposed model. This study contributed to the theoretical and methodological aspects of recommender system study by proposing a methodology that effectively utilizes characteristics of customer's preferences in online reviews.