Product Recommender Systems using Multi-Model Ensemble Techniques

Lee, Yeonjeong; Kim, Kyoung-jae

doi:10.13088/jiis.2013.19.2.039

Cited by 4 publications

(3 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…User classification [4,5] , Preference/Probability calculation [6][7][8] Contents-based filtering [9][10][11] , Kernal function [12] , Weight method [13] Additional information usage Questionnaire [14][15][16] , Previous case [17][18][19] Twitter profile and twit of an user [2] ,…”

Section: Algorithm Improvementmentioning

confidence: 99%

Case Study of Big Data-Based Agri-food Recommendation System According to Types of Customers

Moon¹,

Jang²,

Choe³

et al. 2015

The Journal of Korean Institute of Communications and Informati

View full text Add to dashboard Cite

The Korea Agency of Education, Promotion and Information Service in Food, Agriculture, Forestry and Fisheries launched a public data portal service in January 2015. The service provides customized information for consumers through an agri-food recommendation system built-in portal service. The recommendation system has fallowing characteristics. First, the system can increase recommendation accuracy by using a wide variety of agri-food related data, including SNS opinion mining, consumer's purchase data, climate data, and wholesale price data. Second, the system uses segmentation method based on consumer's lifestyle and megatrends factors to overcome the cold start problem. Third, the system recommends agri-foods to users reflecting various preference contextual factors by using recommendation algorithm, dirichlet-multinomial distribution. In addition, the system provides diverse information related to recommended agri-foods to increase interest in agri-food of service users.

show abstract

Section: Algorithm Improvementmentioning

confidence: 99%

Case Study of Big Data-Based Agri-food Recommendation System According to Types of Customers

Moon¹,

Jang²,

Choe³

et al. 2015

The Journal of Korean Institute of Communications and Informati

View full text Add to dashboard Cite

show abstract

“…Association analysis has been utilized in various fields [10]. Further, there are studies that use both association analysis and social network analysis [11,12,13,14,15,16]. Social network analysis is a quantitative analysis technique that identifies the characteristics of the connection structure and connection state of an object in a group through visual representation [17,18,19,20,21].…”

Section: Association Analysis and Socialmentioning

confidence: 99%

Methodology for Issue-related R&D Keywords Packaging Using Text Mining

Hyun¹,

Shun²,

Kim³

2015

Journal of Internet Computing and Services

View full text Add to dashboard Cite

Considerable research efforts are being directed towards analyzing unstructured data such as text files and log files using commercial and noncommercial analytical tools. In particular, researchers are trying to extract meaningful knowledge through text mining in not only business but also many other areas such as politics, economics, and cultural studies. For instance, several studies have examined national pending issues by analyzing large volumes of text on various social issues. However, it is difficult to provide successful information services that can identify R&D documents on specific national pending issues. While users may specify certain keywords relating to national pending issues, they usually fail to retrieve appropriate R&D information primarily due to discrepancies between these terms and the corresponding terms actually used in the R&D documents. Thus, we need an intermediate logic to overcome these discrepancies, also to identify and package appropriate R&D information on specific national pending issues. To address this requirement, three methodologies are proposed in this study-a hybrid methodology for extracting and integrating keywords pertaining to national pending issues, a methodology for packaging R&D information that corresponds to national pending issues, and a methodology for constructing an associative issue network based on relevant R&D information. Data analysis techniques such as text mining, social network analysis, and association rules mining are utilized for establishing these methodologies. As the experiment result, the keyword enhancement rate by the proposed integration methodology reveals to be about 42.8%. For the second objective, three key analyses were conducted and a number of association rules between national pending issue keywords and R&D keywords were derived. The experiment regarding to the third objective, which is issue clustering based on R&D keywords is still in progress and expected to give tangible results in the future.

show abstract

“…. 이러한 협업필터링 기반 추 천시스템은 Goldberg et al(1992) (Billsus and Pazzani, 1998;Lee and Park, 2007;Bell et al, 2009;Bar et al, 2013;Lee and Kim, 2013 (Bok and Yoo, 2014;Ward and Barker, 2013…”

mentioning

confidence: 99%

A Multimodal Profile Ensemble Approach to Development of Recommender Systems Using Big Data

Kim¹,

Cho²

2015

Journal of Intelligence and Information Systems

View full text Add to dashboard Cite

then neighbors of users are found from the results. One of three ensemble approaches is applied to calculate the similarity. Each ensemble approach uses the similarity of combined profile, the average similarity of each profile, and the weighted average similarity of each profile, respectively. Finally, the products that people among the neighborhood prefer most to are recommended to the target users.For the experiments, we used the demographic data and a very large volume of Web log transaction for 5,000 panel users of a company that is specialized to analyzing ranks of Web sites. R and SAS E-miner was used to implement the proposed recommender system and to conduct the topic analysis using the keyword search, respectively. To evaluate the recommendation performance, we used 60% of data for training and 40% of data for test. The 5-fold cross validation was also conducted to enhance the reliability of our experiments. A widely used combination metric called F1 metric that gives equal weight to both recall and precision was employed for our evaluation. As the results of evaluation, the proposed methodology achieved the significant improvement over the single profile based CF algorithm. In particular, the ensemble approach using weighted average similarity shows the highest performance. That is, the rate of improvement in F1 is 16.9 percent for the ensemble approach using weighted average similarity and 8.1 percent for the ensemble approach using average similarity of each profile. From these results, we conclude that the multimodal profile ensemble approach is a viable solution to the problems encountered when there's a shortage of customer ratings.This study has significance in suggesting what kind of information could we use to create profile in the environment of big data and how could we combine and utilize them effectively. However, our methodology should be further studied to consider for its real-world application. We need to compare the differences in recommendation accuracy by applying the proposed method to different recommendation algorithms and then to identify which combination of them would show the best performance.

show abstract

Product Recommender Systems using Multi-Model Ensemble Techniques

Cited by 4 publications

References 15 publications

Case Study of Big Data-Based Agri-food Recommendation System According to Types of Customers

Case Study of Big Data-Based Agri-food Recommendation System According to Types of Customers

Methodology for Issue-related R&D Keywords Packaging Using Text Mining

A Multimodal Profile Ensemble Approach to Development of Recommender Systems Using Big Data

Contact Info

Product

Resources

About