• Journal of Information Processing Systems
• /
• 제17권2호
• /
• pp.369-384
• /
• 2021

In a personalized product recommendation system, when the amount of log data is large or sparse, the accuracy of model recommendation will be greatly affected. To solve this problem, a personalized product recommendation method using deep factorization machine (DeepFM) to analyze user behavior is proposed. Firstly, the K-means clustering algorithm is used to cluster the original log data from the perspective of similarity to reduce the data dimension. Then, through the DeepFM parameter sharing strategy, the relationship between low- and high-order feature combinations is learned from log data, and the click rate prediction model is constructed. Finally, based on the predicted click-through rate, products are recommended to users in sequence and fed back. The area under the curve (AUC) and Logloss of the proposed method are 0.8834 and 0.0253, respectively, on the Criteo dataset, and 0.7836 and 0.0348 on the KDD2012 Cup dataset, respectively. Compared with other newer recommendation methods, the proposed method can achieve better recommendation effect.

• 산업경영시스템학회지
• /
• 제46권3호
• /
• pp.78-88
• /
• 2023

The increasing number of technology transfers from public research institutes in Korea has led to a growing demand for patent recommendation platforms for SMEs. This is because selecting the right technology for commercialization is a critical factor in business success. This study developed a patent recommendation system that uses technology transfer data from the past 10 years to recommend patents that are suitable for SMEs. The system was developed in three stages. First, an item-based collaborative filtering system was developed to recommend patents based on the similarities between the patents that SMEs have previously transferred. Next, a content-based recommendation system based on TF-IDF was developed to analyze patent names and recommend patents with high similarity. Finally, a hybrid system was developed that combines the strengths of both recommendation systems. The experimental results showed that the hybrid system was able to recommend patents that were both similar and relevant to the SMEs' interests. This suggests that the system can be a valuable tool for SMEs that are looking to acquire new technologies.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제37권3호
• /
• pp.185-193
• /
• 2010

최근 유비쿼터스 시대의 도래와 함께 개인화된 서비스를 제공하기 위한 다양한 서비스 모델들이 제안되어 왔으며, 특히, 사용자에게 개인화된 서비스를 선응적으로 제공하기 위한 다양한 추천 서비스 기법들이 고안되었다. 그러나, 기존의 기법들은 수 많은 데이터를 여과 과정 없이 분석함으로써 추천의 효율성이 떨어지며, 한정된 상황 인지 정보만용 추천 요소로 고려하기 때문에 사용자에게 개인화된 서비스를 제공하기에 적합하지 않다. 본 논문에서는 유비쿼터스 환경에서 사용자의 현재 상황에 가장 적합한 서비스를 제공하는 적응형 추천 서비스 기법을 제안한다. 본 기법은 사용자의 선호도 예측을 위해 누적된 사용자와 장치 간의 상호작용 상황 정보들을 이용하며, 군집 및 협업 필터링 기법을 이용하여 사용자에게 현재 상황에 적응적인 서비스를 추천한다. 군집 기법을 통해 사용자의 현재 위치에 근접한 데이터만을 분석함으로써, 추천의 효율성을 높이며, 협업 필터링을 이용하여 누적된 정보들이 충분하지 않은 상황에서도 정확한 추천을 보장한다. 끝으로, 시뮬레이션을 통해 본 기법의 성능 및 신뢰성을 평가한다.

• 정보과학회 논문지
• /
• 제42권5호
• /
• pp.664-670
• /
• 2015

인터넷과 스마트폰의 발전과 함께 소셜 미디어 공유 사이트인 유튜브도 크게 성장하여 수많은 동영상을 공유하는 사이트가 됐다. 사용자들이 유튜브를 통해 동영상을 공유하면서 소셜 데이터를 만들어내고, 많은 동영상들 중에서 본인의 관심사가 반영된 동영상 추천을 원하게 된다. 본 논문에서는 유튜브 데이터를 이용하여 사용자의 사회적 관계와 유튜브의 특징이 반영된 소셜 카테고리 분류 목록을 기반으로 사용자의 소셜 카테고리를 추출한다. 우리는 좀 더 정확하고 의미있는 추천을 위해 추출된 사용자 소셜 카테고리를 이용한 유튜브 동영상을 추천하는 알고리즘을 제안하였다. 또한 실험을 통해 그 유효성을 검증하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권5호
• /
• pp.2277-2298
• /
• 2019

Adaptive learning in e-learning has garnered researchers' interest. In it, learning resources could be recommended automatically to achieve a personalized learning experience. There are various ways to realize it. One of the realistic ways is adaptive learning path recommendation, in which learning resources are provided according to learners' requirements. This paper summarizes existing works and proposes an innovative approach. Firstly, a learner-centred concept map is created using graph theory based on the features of the learners and concepts. Then, the approach generates a linear concept sequence from the concept map using the proposed traversal algorithm. Finally, Learning Objects (LOs), which are the smallest concrete units that make up a learning path, are organized based on the concept sequences. In order to realize this step, we model it as a multi-objective combinatorial optimization problem, and an improved immune algorithm (IIA) is proposed to solve it. In the experimental stage, a series of simulated experiments are conducted on nine datasets with different levels of complexity. The results show that the proposed algorithm increases the computational efficiency and effectiveness. Moreover, an empirical study is carried out to validate the proposed approach from a pedagogical view. Compared with a self-selection based approach and the other evolutionary algorithm based approaches, the proposed approach produces better outcomes in terms of learners' homework, final exam grades and satisfaction.

• 한국지능시스템학회논문지
• /
• 제16권2호
• /
• pp.164-171
• /
• 2006

최근에 서비스되기 시작한 디지털 멀티미디어 방송은 다양한 종류의 수많은 컨텐츠를 제공하기 때문에 고객은 때로 자신이 선호하는 컨텐츠를 찾는데 많은 시간을 소비한다. 심지어는 선호 컨텐츠를 찾는 동안 이미 방송이 끝날 수도 있다. 이와 같은 문제를 해결하기 위해서는 고객이 필요로 하는 최소 정보만을 추천하기 위한 방법이 필요하다. 본 논문에서는 고객이 시청한 컨텐츠 선호도 전이 확률을 이용하여 고객이 선호하는 컨텐츠를 미리 예측하여 추천하기 위한 알고리즘과 시스템을 제안한다. 제안하는 시스템은 클라이언트 관리자 에이전트, 모니터링 에이전트, 러닝 에이전트, 그리고 추천 에이전트 모듈로 구성된다. 클라이언트 관리자 에이전트는 다른 모듈과 상호 작용을 하면서 조정자 역할을 한다. 모니터링 에이전트는 컨텐츠에 대한 고객의 선호도를 분석하기 위해 고객이 이용했던 usage history 데이터를 수집하기 위한 에이전트이다. 러닝 에이전트는 고객으로부터 수집된 usage history 데이터를 정제하여 시간 변화에 따른 상태 전이 행렬로 모델링하기 위한 에이전트이다. 추천 에이전트는 고객의 상태 전이 행렬로 구성된 모델링 데이터에 본 논문에서 제안하는 선호도 전이 확률 모델을 이용하여 고객이 바로 다음에 선호하게 될 컨텐츠를 추천하기 위한 에이전트이다. 추천 에이전트 모듈에서 컨텐츠에 대한 고객의 선호도 전이 확률을 이용하는 추천 알고리즘을 제안한다. 제안하는 추천 시스템은 무선 인터넷 표준 플랫폼인 WIPI(Wireless Internet Platform for Interoperability) 플랫폼에서 프로토타입 시스템을 설계, 구현하였으며, 실험결과 제안된 선호도 전이 확률 모델의 추천 정확도가 전형적인 방법에 비해 효과적임을 보인다.

• International Journal of Computer Science & Network Security
• /
• 제21권3호
• /
• pp.177-184
• /
• 2021

All over the world, people are affected by many chronic diseases and medical practitioners are working hard to find out the symptoms and remedies for the diseases. Many researchers focus on the feature detection of the disease and trying to get a better health recommendation system. It is necessary to detect the features automatically to provide the most relevant solution for the disease. This research gives the framework of Health Recommendation System (HRS) for identification of relevant and non-redundant features in the dataset for prediction and recommendation of diseases. This system consists of three phases such as Pre-processing, Feature Selection and Performance evaluation. It supports for handling of missing and noisy data using the proposed Imputation of missing data and noise detection based Pre-processing algorithm (IMDNDP). The selection of features from the pre-processed dataset is performed by proposed ensemble-based feature selection using an expert's knowledge (EFS-EK). It is very difficult to detect and monitor the diseases manually and also needs the expertise in the field so that process becomes time consuming. Finally, the prediction and recommendation can be done using Support Vector Machine (SVM) and rule-based approaches.

• Communications for Statistical Applications and Methods
• /
• 제10권1호
• /
• pp.61-77
• /
• 2003

Due to the wide expansion of the internet, people can freely get information what they want with lesser efforts. However without adequate forms or rules to follow, it is getting more and more difficult to get necessary information. Because of seemingly chaotic status of the current web environment, it is sometimes called "Dizzy web" The user should wander from page to page to get necessary information. Therefore we need to construct system which properly recommends appropriate information for general user. The representative research field for this system is called Recommendation System(RS), The collaborative recommendation system is one of the RS. It was known to perform better than the other systems. When we perform the web user modeling or other web-mining tasks, the continuous feedback data is very important and frequently used. In this paper, we propose a collaborative recommendation system which can deal with the continuous feedback data and tried to construct the web page prediction system. We use a sojourn time of a user as continuous feedback data and combine the traditional model-based algorithm framework with the Support Vector Regression technique. In our experiments, we show the accuracy of our system and the computing time of page prediction compared with Pearson's correlation algorithm.algorithm.

• 대한산업공학회지
• /
• 제39권2호
• /
• pp.135-142
• /
• 2013

In B2B e-Marketplace for free gifts and goods, product-mix recommendation is provided frequently by analysing customer logs and/or performing collaborative and rules-based filtering. This study proposes a new process that encompasses the genetic algorithm and key working processes of B2B e-marketplace based on the previous cooperate client order data. Efficiency and accuracy of the proposed system have been confirmed by cross-confirmation of accumulated data in the e-marketplace. The system can provide better opportunities for manufactures and suppliers to select optimized product-mix without time consuming trials and errors in their B2B e-marketplace networks.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 컴퓨터소사이어티 추계학술대회논문집
• /
• pp.175-178
• /
• 2003

Two defects have been pointed out in existing user-based collaborative filtering such as sparsity and scalability, and the research has been also made progress, which tries to improve these defects using item-based collaborative filtering. Actually there were many results, but the problem of sparsity still remains because of being based on an explicit data. In addition, the issue has been pointed out. which attributes of item arenot reflected in the recommendation. This paper suggests a recommendation method using nave Bayesian algorithm in hybrid user and item-based collaborative filtering to improve above-mentioned defects of existing item-based collaborative filtering. This method generates a similarity table for each user and item, then it improves the accuracy of prediction and recommendation item using naive Bayesianalgorithm. It was compared and evaluated with existing item-based collaborative filtering technique to estimate the accuracy.