Basics of qualitative research: Grounded theory procedures and techniques

A central problem in machine learning is identifying a representative set of features from which to construct a classification model for a particular task. This thesis addresses the problem of feature selection for machine learning through a correlation based approach. The central hypothesis is that good feature sets contain features that are highly correlated with the class, yet uncorrelated with each other. A feature evaluation formula, based on ideas from test theory, provides an operational definition of this hypothesis. CFS (Correlation based Feature Selection) is an algorithm that couples this evaluation formula with an appropriate correlation measure and a heuristic search strategy. CFS was evaluated by experiments on artificial and natural datasets. Three machine learning algorithms were used: C4.5 (a decision tree learner), IB1 (an instance based learner), and naive Bayes. Experiments on artificial datasets showed that CFS quickly identifies and screens irrelevant, redundant, and noisy features, and identifies relevant features as long as their relevance does not strongly depend on other features. On natural domains, CFS typically eliminated well over half the features. In most cases, classification accuracy using the reduced feature set equaled or bettered accuracy using the complete feature set. Feature selection degraded machine learning performance in cases where some features were eliminated which were highly predictive of very small areas of the instance space. Further experiments compared CFS with a wrapper—a well known approach to feature selection that employs the target learning algorithm to evaluate feature sets. In many cases CFS gave comparable results to the wrapper, and in general, outperformed the wrapper on small datasets. CFS executes many times faster than the wrapper, which allows it to scale to larger datasets. Two methods of extending CFS to handle feature interaction are presented and experimentally evaluated. The first considers pairs of features and the second incorporates iii feature weights calculated by the RELIEF algorithm. Experiments on artificial domains showed that both methods were able to identify interacting features. On natural domains, the pairwise method gave more reliable results than using weights provided by RELIEF.

/pdf/correlation-based-feature-selection-for-machine-learning-578rirmvg5.pdf

Correlation-based Feature Selection for Machine Learning

We have witnessed great interest and a wealth of promise in content-based image retrieval as an emerging technology. While the last decade laid foundation to such promise, it also paved the way for a large number of new techniques and systems, got many new people involved, and triggered stronger association of weakly related fields. In this article, we survey almost 300 key theoretical and empirical contributions in the current decade related to image retrieval and automatic image annotation, and in the process discuss the spawning of related subfields. We also discuss significant challenges involved in the adaptation of existing image retrieval techniques to build systems that can be useful in the real world. In retrospect of what has been achieved so far, we also conjecture what the future may hold for image retrieval research.

/pdf/image-retrieval-ideas-influences-and-trends-of-the-new-age-51whw4w7su.pdf

Image retrieval: Ideas, influences, and trends of the new age

The Visual Display of Quantitative Information

We analyzed transaction logs containing 51,473 queries posed by 18,113 users of Excite, a major Internet search service. We provide data on: (i) sessions — changes in queries during a session, number of pages viewed, and use of relevance feedback; (ii) queries — the number of search terms, and the use of logic and modifiers; and (iii) terms — their rank/frequency distribution and the most highly used search terms. We then shift the focus of analysis from the query to the user to gain insight to the characteristics of the Web user. With these characteristics as a basis, we then conducted a failure analysis, identifying trends among user mistakes. We conclude with a summary of findings and a discussion of the implications of these findings. # 2000 Elsevier Science Ltd. All rights reserved.

/pdf/real-life-real-users-and-real-needs-a-study-and-analysis-of-5bl07cx90t.pdf

Real life, real users, and real needs: a study and analysis of user queries on the web

The Waikato Environment for Knowledge Analysis (Weka) is a comprehensive suite of Java class libraries that implement many state-of-the-art machine learning and data mining algorithms. Weka is freely available on the World-Wide Web and accompanies a new text on data mining [1] which documents and fully explains all the algorithms it contains. Applications written using the Weka class libraries can be run on any computer with a Web browsing capability; this allows users to apply machine learning techniques to their own data regardless of computer platform.

/pdf/weka-practical-machine-learning-tools-and-techniques-with-4t328z5hr1.pdf

Weka: Practical machine learning tools and techniques with Java implementations

Music is traditionally retrieved by title, composer or subject classification. It is possible, with current technology, to retrieve music from a database on the basis of a few notes sung or hummed into a microphone. This paper describes the implementation of such a system, and discusses several issues pertaining to music retrieval. We first describe an interface that transcribes acoustic input into standard music notation. We then analyze string matching requirements for ranked retrieval of music and present the results of an experiment which tests how accurately people sing well known melodies. The performance of several string matching criteria are analyzed using two folk song databases. Finally, we describe a prototype system which has been developed for retrieval of tunes from acoustic input.

/pdf/towards-the-digital-music-library-tune-retrieval-from-1n7dze0v0e.pdf

Towards the digital music library: tune retrieval from acoustic input

As experimental digital library testbeds gain wider acceptance and develop significant user bases, it becomes important to investigate the ways in which users interact with the systems in practice. Transaction logs are one source of usage information, and the information on user behavior can be culled from them both automatically (through calculation of summary statistics) and manually (by examining query strings for semantic clues on search motivations and searching strategy). We have conducted a transaction log analysis on user activity in the Computer Science Technical Reports Collection of the New Zealand Digital Library, and report insights gained and identify resulting search interface design issues. Specifically, we present the user demographics available with our library, discuss the use of operators and search options in queries, and examine patterns in query construction and refinement. We also describe common mistakes in searching, and examine the distribution of query terms appearing in the logs.

A transaction log analysis of a digital library

At present, music digital library systems are being developed based on anecdotal evidence of user needs, intuitive feelings for user information seeking behavior, and a priori assumptions of typical usage scenarios. Emphasis has been placed on basic research into music document representation, efficient searching, and audio-based searching, rather than on exploring the music information needs or information behavior of a target user group. This paper focuses on eliciting the native music information strategies employed by people searching for popular music (that is, music sought for recreational or enjoyment purposes rather than to support a serious or scientific exploration of some aspect of music). To this end, we conducted an ethnographic study of the searching/browsing techniques employed by people in the researchers local communities, as they use two common sources of music: the public library and music stores. We argue that the insights provided by this type of study can inform the development of searching/browsing support for music digital libraries.

http://web.mit.edu/bentley/www/mobile/papers/cunningham2003.pdf

An ethnographic study of music information seeking: implications for the design of a music digital library

This paper investigates the information seeking of humanities academics and scholars using digital libraries. It furthers existing work by updating our knowledge of the information seeking techniques used by humanities scholars, where the current work predates the wide availability of the Internet. We also report some of the patterns observed in query and term usage by humanities scholars, and relate this to the patterns they report in their own information seeking and the problems that they encounter. This insight is used to reveal the current gap between the skills of information seekers and the technologies that they use. Searches for ‘discipline terms' prove to be particularly problematic.

/pdf/information-seeking-by-humanities-scholars-2ywrmyejze.pdf

Sally Jo Cunningham

Papers

Weka: Practical machine learning tools and techniques with Java implementations

Towards the digital music library: tune retrieval from acoustic input

A transaction log analysis of a digital library

An ethnographic study of music information seeking: implications for the design of a music digital library

Information seeking by humanities scholars