• Journal of the Korea Society of Computer and Information
• /
• v.27 no.9
• /
• pp.149-155
• /
• 2022

The popularity of e-commerce systems on the Internet is increasing day by day, and the recommendation system, as a core function of these systems, greatly reduces the effort to search for desired products by recommending products that customers may prefer. The collaborative filtering technique is a recommendation algorithm that has been successfully implemented in many commercial systems, but despite its popularity and usefulness in academia, the memory-based implementation has inaccuracies in its reference neighbor. To solve this problem, this study proposes a new time-aware collaborative filtering technique that integrates and utilizes the neighbors of each item and each user, weighting the recent similarity more than the past similarity with them, and reflecting it in the recommendation list decision. Through the experimental evaluation, it was confirmed that the proposed method showed superior performance in terms of prediction accuracy than other existing methods.

• The Journal of Korean Association of Computer Education
• /
• v.19 no.4
• /
• pp.59-66
• /
• 2016

It has been studied to reflect the number of co-rated items for solving data sparsity problem in collaborative filtering systems. A well-known method of Jaccard index allowed performance improvement, when combined with previous similarity measures. However, the degree of performance improvement when combined with existing similarity measures in various data environments are seldom analyzed, which is the objective of this study. Jaccard index as a sole similarity measure yielded much higher prediction quality than traditional measures and very high recommendation quality in a sparse dataset. In general, previous similarity measures combined with Jaccard index improved performance regardless of dataset characteristics. Especially, cosine similarity achieved the highest improvement in sparse datasets, while similarity of Mean Squared Difference degraded prediction quality in denser sets. Therefore, one needs to consider characteristics of data environment and similarity measures before combining Jaccard index for similarity use.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.12
• /
• pp.259-266
• /
• 2023

In this paper, we propose an algorithm that can improve the accuracy performance of collaborative filtering using attribute-based opinion mining (ABOM). For the experiment, a total of 1,227 online consumer review data about smartphone apps from domestic smartphone users were used for analysis. After morpheme analysis using the KKMA (Kkokkoma) analyzer and emotional word analysis using KOSAC, attribute extraction is performed using LDA topic modeling, and the topic modeling results for each weighted review are used to add up the ratings of collaborative filtering and the sentiment score. MAE, MAPE, and RMSE, which are statistical model performance evaluations that calculate the average accuracy error, were used. Through experiments, we predicted the accuracy of online customers' app ratings (APP_Score) by combining traditional collaborative filtering among the recommendation algorithms and the attribute-based opinion mining (ABOM) technique, which combines LDA attribute extraction and sentiment analysis. As a result of the analysis, it was found that the prediction accuracy of ratings using attribute-based opinion mining CF was better than that of ratings implementing traditional collaborative filtering.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.2 no.2
• /
• pp.115-119
• /
• 2001

The design and manufacture of injection molded polymeric parts with desired properties is a costly process dominated by empiricism, including the repeated modification of actual tooling. This paper presents an knowledge-based synthesis system which can predict the mechanical performance of a molded product and diagnose the design before the actual mold is machined. The knowledge-based system synergistically combines a rule-based system with CAE programs. Hueristic know]edge of injection molding. flow simulation, and mechanical performance prediction is formalized as rules of an expert consultation system. The expert system interprets the analytical results of the process simulation, predicts the performance, evaluates the design and generates recommendation for optimal design alternatives.

• The Journal of Korean Association of Computer Education
• /
• v.21 no.5
• /
• pp.61-68
• /
• 2018

Collaborative filtering is a fundamental technique implemented in many commercial recommender systems and provides a successful service to online users. This technique recommends items by referring to other users who have similar rating records to the current user. Hence, similarity measures critically affect the system performance. This study addresses problems of previous similarity measures and suggests a new similarity measure. The proposed measure reflects the subjectivity or vagueness of user ratings and the users' rating behavior by using fuzzy logic. We conduct experimental studies for performance evaluation, whose results show that the proposed measure demonstrates outstanding performance improvements in terms of prediction accuracy and recommendation accuracy.

• Korean Business Review
• /
• v.20 no.2
• /
• pp.105-118
• /
• 2007

Recommender system is the tool for E-commerce company based on the internet for increasing their sales ratio in the market. Recommender system suggests the list of items which night be wanted by customers. This list generated by the result of customers' preference prediction through the prediction algorithm automatically. Recommender system will be able to offer not only the important information for marketing strategy but also reduce the cost of customers' information retrieval trough the analysis of customers' purchase patterns and features. But there are several problems like as the extension of the users and items scales and if the recommendation to customers generated by unreliable recommender system makes the customer royalty to the system to weaken. In this study, we propose the criterion for pre-evaluation on the prediction performance only using the preference ratings on the items which are rated by customers before prediction process and we study the features of customers who are classified through this classification criterion.

• Ocean Systems Engineering
• /
• v.7 no.4
• /
• pp.319-328
• /
• 2017

Higher prediction efficacy is a very challenging task in any field of engineering. Due to global warming, there is a considerable increase in the global sea level. Through this work, an attempt has been made to find the sea level variability due to climate change impact at Haldia Port, India. Different statistical downscaling techniques are available and through this paper authors are intending to compare and illustrate the performances of three regression models. The models: Wavelet Neural Network (WNN), Minimax Probability Machine Regression (MPMR), Feed-Forward Neural Network (FFNN) are used for projecting the sea level variability due to climate change at Haldia Port, India. Model performance indices like PI, RMSE, NSE, MAPE, RSR etc were evaluated to get a clear picture on the model accuracy. All the indices are pointing towards the outperformance of WNN in projecting the sea level variability. The findings suggest a strong recommendation for ensembled models especially wavelet decomposed neural network to improve projecting efficiency in any time series modeling.

• Journal of Information Technology Services
• /
• v.21 no.4
• /
• pp.75-90
• /
• 2022

From the past to the present, reviews have had much influence on consumers' purchasing decisions. Companies are making various efforts, such as introducing a review incentive system to increase the number of reviews. Recently, as various types of reviews can be left, reviews have begun to be recognized as interesting new content. This way, reviews have become essential in creating loyal customers. Therefore, research and utilization of reviews are being actively conducted. Some studies analyze reviews to discover customers' needs, studies that upgrade recommendation systems using reviews, and studies that analyze consumers' emotions and attitudes through reviews. However, research that predicts the future using reviews is insufficient. This study used a dataset consisting of two reviews written in pairs with differences in usage periods. In this study, the direction of consumer product evaluation is predicted using KoBERT, which shows excellent performance in Text Deep Learning. We used 7,233 reviews collected to demonstrate the excellence of the proposed model. As a result, the proposed model using the review text and the star rating showed excellent performance compared to the baseline that follows the majority voting.

• Journal of KIISE:Software and Applications
• /
• v.35 no.10
• /
• pp.608-615
• /
• 2008

In Knowledge Q&A services where information is created by unspecified users, document quality is an important factor of user satisfaction with search results. Previous work on quality prediction of Knowledge Q&A documents evaluate the quality of documents by using non-textual information, such as click counts and recommendation counts, and focus on enhancing retrieval performance by incorporating the quality measure into retrieval model. Although the non-textual information used in previous work was proven to be useful by experiments, data sparseness problem may occur when predicting the quality of newly created documents with such information. To solve data sparseness problem of non-textual features, this paper proposes new features for document quality prediction, namely text-confidence features, which indicate how trustworthy the content of a document is. The proposed features, extracted directly from the document content, are stable against data sparseness problem, compared to non-textual features that indirectly require participation of service users in order to be collected. Experiments conducted on real world Knowledge Q&A documents suggests that text-confidence features show performance comparable to the non-textual features. We believe the proposed features can be utilized as effective features for document quality prediction and improve the performance of Knowledge Q&A services in the future.

• Knowledge Management Research
• /
• v.22 no.4
• /
• pp.27-40
• /
• 2021

Click-Through Rate (CTR) prediction is a key function that determines the ranking of candidate items in the recommendation system and recommends high-ranking items to reduce customer information overload and achieve profit maximization through sales promotion. The fields of natural language processing and image classification are achieving remarkable growth through the use of deep neural networks. Recently, a transformer model based on an attention mechanism, differentiated from the mainstream models in the fields of natural language processing and image classification, has been proposed to achieve state-of-the-art in this field. In this study, we present a method for improving the performance of a transformer model for CTR prediction. In order to analyze the effect of discrete and categorical CTR data characteristics different from natural language and image data on performance, experiments on embedding regularization and transformer normalization are performed. According to the experimental results, it was confirmed that the prediction performance of the transformer was significantly improved when the L2 generalization was applied in the embedding process for CTR data input processing and when batch normalization was applied instead of layer normalization, which is the default regularization method, to the transformer model.