Earticle

본 연구에서는 관광지 추천 알고리즘의 예측 정확도 결정요인을 알고리즘이 보여주는 시스템 차원이 아니라 합리적 의사결정 과정에서 핵심 요소로 작용하는 동시에 경쟁관계를 보이는 두 요소인 선호와 거리 간 위계적 순위 관점에서 고찰하였다. 소비자에게 전달하는 제품의 주요 가치 가운데 관광 상품은 기능적 가치와 상징적 가치를 우위하는 경험적 가치에 근간을 두고 있다. 경험적 가치를 근본으로 하는 관광지 선정 과정에는 선호가 거리를 우위하는 위계적 순위를 보인다. 따라서 선호와 거리를 동시에 고려하되 거리보다 선호에 더 큰 가중치를 둔 추천 알고리즘이 단순 히 선호나 거리만을 중심으로 하는 추천 알고리즘보다 예측력이 우수한 것으로 나타났다. 실증 분석을 위한 표본은 강원도 춘천 관광지에 대한 평 점(별점)과 평점 횟수, 후기와 후기 횟수 등으로 구성된 데이터셋을 사용하였다. 실증분석에 사용된 자료는 네이버 지도를 통해 수집하였고, 추천 알고리즘의 예측력 비교를 위해 정밀도(Precision)와 재현율(Recall)을 보여주는 P-R 곡선을 분석도구로 사용하였다. 실증분석 결과 선호와 거리 를 순차적으로 적용한 추천 알고리즘이 선호 또는 거리만을 고려한 추천 알고리즘보다 P-R곡선이 기준치인 1 값에 인접한 채 하락하는 것으로 나 타났다(가설 2지지). 이 연구에서 발견된 사실은 개별 추천 알고리즘의 기술적 우위를 벗어나 추천 알고리즘의 예측력은 선호와 거리라는 경쟁하 는 요소 간 위계적 순위에 따라 결정된다는 논리를 지지해준다. 따라서 정부와 지자체 그리고 관련 기업은 선호와 거리를 모두 고려하여 관광지를 개발하고, 선호와 거리를 순차적으로 고려한 비즈니스 마케팅 전략을 수립해야 한다.

This study examines the factors that determine the predictive accuracy of the tourist recommendation algorithm not at the system level of the algorithm, but in terms of the hierarchical ranking between preference and distance, which are two competing elements that act as key elements in the rational decision-making process. Among the main values that products want to deliver to consumers, tourism products are based on experiential values that predominate over functional and symbolic values. In the process of selecting a tourist destination based on experiential value, preference shows a hierarchical ranking that takes precedence over distance. Therefore, it was found that the recommendation algorithm that considers both preference and distance at the same time, but puts more weight on preference than distance, has better predictive power than recommendation algorithms that simply focus only on preference or distance. For the sample for empirical analysis, a data set consisting of ratings (star ratings) for tourist attractions in Chuncheon, Gangwon-do, the number of ratings, and the number of reviews and reviews was used. The data used in the empirical analysis were collected through Naver Map, and the P-R curve showing precision and recall was used as an analysis tool to compare the predictive power of the recommendation algorithm. As a result of the empirical analysis, it was found that the recommendation algorithmthat sequentially applied preference and distance decreased the P-R curve while remaining close to the reference value of 1, rather than the recommendation algorithm that only considered preference or distance (Hypothesis 2 supported). The findings in this study go beyond the technical superiority of individual recommendation algorithms and support the logic that the predictive power of recommendation algorithms is determined by the hierarchical ranking of competing factors such as preference and distance. Therefore, the government, local governments, and related companies should develop tourismpolicies considering both preference and distance, and establish businessmarketing strategies that sequentially consider the significance of preference and distance.

오피니언 마이닝은 기존의 데이터 마이닝 기술을 활용하여 웹 상에 개재된 블로그, 상품평등에 나타난 저자의 의견을 추출하는 분야로써 텍스트의 주제를 판단하는 것이 아닌 주제에 대한 저자의 태도를 판단하는 기술이다. 본 논문에서는 오피니언 마이닝 알고리즘과 공개된 음성인식 API을 사용하여 텍스트가 아닌 음성의 대한 데이터의 감정을 판단하기 위해 제안했다. 이 시스템은 공개된 Google Voice Recognition API와 주제어와 관련된 순위화 알고리즘, 개선된 극성 판단 알고리즘을 통하여 설계하고, 이를 바탕으로 음성인식 인터뷰 모델을 구현한다.

The opinion mining is that to use the existing data mining technology also uploaded blog to web, to use product comment, the opinion mining can extract the author's opinion therefore it not judge text's subject, only judge subject's emotion. In this paper, published opinion mining algorithms and the text using speech recognition API for non-voice data to judge the emotions suggested. The system is open and the Subject associated with Google Voice Recognition API sunwihwa algorithm, the algorithm determines the polarity through improved design, based on this interview, speech recognition, which implements the model.

Most information in Web 2.0 is made by users and classified by tags assigned by users. Tag-related services and research are focused on work such as automatic tagging and tag-cloud composition; however, classifying media resources and information according to tags and providing the results to users is not still up to the mark. In this paper, image resources and their tag information scattered in the web are collected and a tag-pair weight matrix is created, according to the relations and semantic similarities between tags. To overcome the problems of the existing system, a tag-pair weight matrix-based tag clustering (TBTC) algorithm was proposed to find highly related tags. The threshold used for clustering in this algorithm was studied, and an optimal threshold with high cluster cohesion was determined. Finally, as an experiment, 500 images with the keyword 'tomato' were searched from the Flickr website and highly related tags were derived from the proposed algorithm. The results of this research were examined and compared with the results of existing studies. It was found that the proposed research showed more advanced accuracy and precision than earlier methods.

Objectives: The purpose of this study was to investigate the effects of query expansion algorithms for MEDLINE retrieval within a pseudo-relevance feedback framework. Methods: A number of query expansion algorithms were tested using various term ranking formulas, focusing on query expansion based on pseudo-relevance feedback. The OHSUMED test collection,which is a subset of the MEDLINE database, was used as a test corpus. Various ranking algorithms were tested in combination with different term re-weighting algorithms. Results: Our comprehensive evaluation showed that the local context analysis ranking algorithm, when used in combination with one of the reweighting algorithms ? Rocchio, the probabilistic model, and our variants ? significantly outperformed other algorithm combinations by up to 12% (paired t-test; p < 0.05). In a pseudo-relevance feedback framework, effective query expansion would be achieved by the careful consideration of term ranking and re-weighting algorithm pairs, at least in the context of the OHSUMED corpus. Conclusions: Comparative experiments on term ranking algorithms were performed in the context of a subset of MEDLINE documents. With medical documents, local context analysis, which uses co-occurrence with all query terms, significantly outperformed various term ranking methods based on both frequency and distribution analyses. Furthermore, the results of the experiments demonstrated that the term rank-based re-weighting method contributed to a remarkable improvement in mean average precision.

텍스트 순위 알고리즘은 키워드 추출을 위한 대표적인 방법이며 그 중요성이 강조되고 있다. 본 논문에서는 텍스트 랭킹 알고리즘에서 대표적으로 사용되는 TF-IDF, SMART, INQUERY, CCA 알고리즘이 적용된 최근 연구와 실험해비교한다. 먼저, 각 알고리즘을 설명한 후 뉴스와 트위터 데이터를 기반으로 알고리즘의 성능을 분석한다. 실험 결과에 따르면 네 가지 알고리즘 모두 뉴스 데이터에서 특정 단어의 추출 성능이 좋다는 것을 알 수 있다. 그러나 Twitter의 경우 CCA는 특정 단어를 추출하는 최고의 성능을 가지며 INQUERY는 가장 낮은 성능을 보여준다. 또한 6 가지 비교 메트릭을 통해 알고리즘의 정확성을 분석한다. 실험 결과 CCA가 뉴스 데이터에서 최고의 정확도를 보여주고, 트위터의 경우 TF-IDF와 CCA는 비슷한 성능을 보이며 높은 정확도를 보인다.

The text ranking algorithm is a representative method for keyword extraction, and its importance is emphasized highly. In this paper, we compare the performance of recent research and experiments with TF-IDF, SMART, INQUERY and CCA algorithms, which are used in text ranking algorithm.. After explaining each algorithm, we compare the performance of each algorithm based on the data collected from news and Twitter. Experimental results show that all of four algorithms can extract specific words from news data equally. However, in the case of Twitter, CCA has the best performance to extract specific words, and INQUERY shows the worst performance. We also analyze the accuracy of the algorithm through six comparison metrics. The experimental results present that CCA shows the best accuracy in the news data. In case of Twitter, TF-IDF and CCA show similar performance and demonstrate good performance.

This article performs a detailed data scrutiny on a chronic kidney disease (CKD) dataset to select efficient instances and relevant features. Data relevancy is investigated using feature extraction, hybrid outlier detection, and handling of missing values. Data instances that do not influence the target are removed using data envelopment analysis to enable reduction of rows. Column reduction is achieved by ranking the attributes through feature selection methodologies, namely, extra-trees classifier, recursive feature elimination, chi-squared test, analysis of variance, and mutual information. These methodologies are ranked via Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) using weight optimization to identify the optimal features for model building from the CKD dataset to facilitate better prediction while diagnosing the severity of the disease. An efficient hybrid ensemble and novel similarity-based classifiers are built using the pruned dataset, and the results are thereafter compared with random forest, AdaBoost, naive Bayes, k-nearest neighbors, and support vector machines. The hybrid ensemble classifier yields a better prediction accuracy of 98.31% for the features selected by extra tree classifier (ETC), which is ranked as the best by TOPSIS.

블로그는 개인의 생각을 온라인상의 문서인 포스트로 남기는 일종의 개인 홈페이지이다. 웹 문서와 포스트는 다른 특징을 가지고 있기 때문에 웹 문서 랭킹 알고리즘을 포스트에 적용하기에는 문제점이 존재한다. 본 논문에서는 포스트의 랭킹을 위한 포스트-블로거 그래프를 정의하고, 기존의 웹 문서 랭킹 알고리즘을 포스트-블로거 그래프에 적합하게 변형한 포스트 랭킹 알고리즘을 제안한다. 그리고 실제 블로그 데이터를 통하여 포스트 랭킹 알고리즘의 성능을 분석하고, 이를 바탕으로 블로그에 적합한 포스트 랭킹 알고리즘을 선별한다.