Term Selection Methods Research Articles

텍스트 마이닝 기법을 이용한 연관용어 선정에 관한 실험적 연구

이 연구에서는 전체 문헌집단으로부터 초기 질의어에 대한 연관용어 선정 시 사용할 수 있는 최적의 기법을 찾기 위해 연관규칙 마이닝과 용어 클러스터링 기법을 이용하여 연관용어 선정 실험을 수행하였다. 연관규칙 마이닝 기법에서는 Apriori 알고리즘을 사용하였으며, 용어 클러스터링 기법에서는 연관성 척도로 GSS 계수, 자카드계수, 코사인계수, 소칼 & 스니스 5, 상호정보량을 사용하였다. 성능평가 척도로는 연관용어 정확률과 연관용어 일치율을 사용하였으며, 실험결과 Apriori 알고리즘과 GSS 계수가 가장 좋은 성능을 나타냈다.

Unbiased variable selection for classification trees with multivariate responses

A variable selection scheme is proposed for constructing multivariate classification trees. It utilizes conditional independence test derived from hierarchical loglinear model for three-way contingency table to control selection bias. Furthermore, it is compared with some existing selection methods in terms of selection power. Simulation results show that our method is unbiased and has better selection power.

Feature Selection for Classification of SELDI-TOF-MS Proteomic Profiles

Proteomic peptide profiling is an emerging technology harbouring great expectations to enable early detection, enhance diagnosis and more clearly define prognosis of many diseases. Although previous research work has illustrated the ability of proteomic data to discriminate between cases and controls, significantly less attention has been paid to the analysis of feature selection strategies that enable learning of such predictive models. Feature selection, in addition to classification, plays an important role in successful identification of proteomic biomarker panels. We present a new, efficient, multivariate feature selection strategy that extracts useful feature panels directly from the high-throughput spectra. The strategy takes advantage of the characteristics of surface-enhanced laser desorption/ionisation time-of-flight mass spectrometry (SELDI-TOF-MS) profiles and enhances widely used univariate feature selection strategies with a heuristic based on multivariate de-correlation filtering. We analyse and compare two versions of the method: one in which all feature pairs must adhere to a maximum allowed correlation (MAC) threshold, and another in which the feature panel is built greedily by deciding among best univariate features at different MAC levels. The analysis and comparison of feature selection strategies was carried out experimentally on the pancreatic cancer dataset with 57 cancers and 59 controls from the University of Pittsburgh Cancer Institute, Pittsburgh, Pennsylvania, USA. The analysis was conducted in both the whole-profile and peak-only modes. The results clearly show the benefit of the new strategy over univariate feature selection methods in terms of improved classification performance. Understanding the characteristics of the spectra allows us to better assess the relative importance of potential features in the diagnosis of cancer. Incorporation of these characteristics into feature selection strategies often leads to a more efficient data analysis as well as improved classification performance.

Building Minority Language Corpora by Learning to Generate Web Search Queries

The Web is a source of valuable information, but the process of collecting, organizing, and effectively utilizing the resources it contains is difficult. We describe CorpusBuilder, an approach for automatically generating Web search queries for collecting documents matching a minority concept. The concept used for this paper is that of text documents belonging to a minority natural language on the Web. Individual documents are automatically labeled as relevant or nonrelevant using a language filter, and the feedback is used to learn what query lengths and inclusion/exclusion term-selection methods are helpful for finding previously unseen documents in the target language. Our system learns to select good query terms using a variety of term scoring methods. Using odds ratio scores calculated over the documents acquired was one of the most consistently accurate query-generation methods. To reduce the number of estimated parameters, we parameterize the query length using a Gamma distribution and present empirical results with learning methods that vary the time horizon used when learning from the results of past queries. We find that our system performs well whether we initialize it with a whole document or with a handful of words elicited from a user. Experiments applying the same approach to multiple languages are also presented showing that our approach generalizes well across several languages regardless of the initial conditions.

Bootstrapping for efficient handwritten digit recognition

In this paper we present two algorithms for selecting prototypes from the given training data set. Here, we employ the bootstrap technique to preprocess the data. We compare the proposed algorithms with the condensed nearest-neighbor algorithm which is order dependent and a genetic-algorithm-based prototype selection scheme which is order independent. Algorithms proposed in this paper are found to be better than the condensed nearest neighbor and prototype selection methods in terms of classification accuracy.

Code Growth, Explicitly Defined Introns, and Alternative Selection Schemes

Previous work on introns and code growth in genetic programming is expanded on and tested experimentally. Explicitly defined introns are introduced to tree-based representations as an aid to measuring and evaluating intron behavior. Although it is shown that introns do create code growth, they are not its only cause. Removing introns merely decreases the growth rate; it does not eliminate it. By systematically negating various forms of intron behavior, a deeper understanding of the causes of code growth is obtained, leading to the development of a system that keeps unnecessary bloat to a minimum. Alternative selection schemes and recombination operators are examined and improvements demonstrated over the standard selection methods in terms of both performance and parsimony.

The Impact on Utility, Race, and Gender Using Three Standard Methods of Scoring Selection Examinations

This study examined the utility implications of three commonly used personnel selection methods in terms of social and economic equity. The comparative analysis stemmed from the controversial issues of weighing social policy as identified by EEOC standards and affirmative action guidelines against an employer's right to hire the most productive job applicants. How the social equity and productivity issues are addressed, according to an organization's desired outcomes, can help determine the type of selection procedure adopted. In terms of productivity, the top-down raw score approach was by far the best, followed by the minimum cutoff procedure and the within-group standardized score method produced the greatest selection of minority groups, followed by the minimum cutoff procedure and the top-down raw score approach. The top-down raw score approach clearly showed a disparate impact on minority groups, while the within-group standardized scoring method demonstrated adverse impact on the white group, indicating a potential for reverse discrimination. The results from comparing these selection strategies in terms of the gain in productivity and in social equity are of enormous assistance when making a final decision as to the appropriate selection strategy.