A person seeking someone else's attention is normally able to quickly assess how interruptible they are. This assessment allows for behavior we perceive as natural, socially appropriate, or simply polite. On the other hand, today's computer systems are almost entirely oblivious to the human world they operate in, and typically have no way to take into account the interruptibility of the user. This paper presents a Wizard of Oz study exploring whether, and how, robust sensor-based predictions of interruptibility might be constructed, which sensors might be most useful to such predictions, and how simple such sensors might be. The study simulates a range of possible sensors through human coding of audio and video recordings. Experience sampling is used to simultaneously collect randomly distributed self-reports of interruptibility. Based on these simulated sensors, we construct statistical models predicting human interruptibility and compare their predictions with the collected self-report data. The results of these models, although covering a demographically limited sample, are very promising, with the overall accuracy of several models reaching about 78%. Additionally, a model tuned to avoiding unwanted interruptions does so for 90% of its predictions, while retaining 75% overall accuracy.

This paper introduces concepts and algorithms of feature selection, surveys existing feature selection algorithms for classification and clustering, groups and compares different algorithms with a categorizing framework based on search strategies, evaluation criteria, and data mining tasks, reveals unattempted combinations, and provides guidelines in selecting feature selection algorithms. With the categorizing framework, we continue our efforts toward building an integrated system for intelligent feature selection. A unifying platform is proposed as an intermediate step. An illustrative example is presented to show how existing feature selection algorithms can be integrated into a meta algorithm that can take advantage of individual algorithms. An added advantage of doing so is to help a user employ a suitable algorithm without knowing details of each algorithm. Some real-world applications are included to demonstrate the use of feature selection in data mining. We conclude this work by identifying trends and challenges of feature selection research and development.

Feature selection is applied to reduce the number of features in many applications where data has hundreds or thousands of features. Existing feature selection methods mainly focus on finding relevant features. In this paper, we show that feature relevance alone is insufficient for efficient feature selection of high-dimensional data. We define feature redundancy and propose to perform explicit redundancy analysis in feature selection. A new framework is introduced that decouples relevance analysis and redundancy analysis. We develop a correlation-based method for relevance and redundancy analysis, and conduct an empirical study of its efficiency and effectiveness comparing with representative methods.

We present a comparative evaluation of automatic classification of a sound database containing more than six hundred drum sounds (kick, snare, hihat, toms and cymbals). A preliminary set of fifty descriptors has been refined with the help of different techniques and some final reduced sets including around twenty features have been selected as the most relevant. We have then tested different classification techniques (instance-based, statistical-based, and tree-based) using ten-fold cross-validation. Three levels of taxonomic classification have been tested: membranes versus plates (super-category level), kick vs. snare vs. hihat vs. toms vs. cymbals (basic level), and some basic classes (kick and snare) plus some sub-classes --i.e. ride, crash, open-hihat, closed hihat, high-tom, medium-tom, low-tom- (sub-category level). Very high hit-rates have been achieved (99%, 97%, and 90% respectively) with several of the tested techniques.

In gene expression microarray data analysis, selecting a small number of discriminative genes from thousands of genes is an important problem for accurate classification of diseases or phenotypes. The problem becomes particularly challenging due to the large number of features (genes) and small sample size. Traditional gene selection methods often select the top-ranked genes according to their individual discriminative power without handling the high degree of redundancy among the genes. Latest research shows that removing redundant genes among selected ones can achieve a better representation of the characteristics of the targeted phenotypes and lead to improved classification accuracy. Hence, we study in this paper the relationship between feature relevance and redundancy and propose an efficient method that can effectively remove redundant genes. The efficiency and effectiveness of our method in comparison with representative methods has been demonstrated through an empirical study using public microarray data sets.

Interactions are patterns between several attributes in data that cannot be inferred from any subset of these attributes. While mutual information is a well-established approach to evaluating the interactions between two attributes, we surveyed its generalizations as to quantify interactions between several attributes. We have chosen McGill’s interaction information, which has been independently rediscovered a number of times under various names in various disciplines, because of its many intuitively appealing properties. We apply interaction information to visually present the most important interactions of the data. Visualization of interactions has provided insight into the structure of data on a number of domains, identifying redundant attributes and opportunities for constructing new features, discovering unexpected regularities in data, and have helped during construction of predictive models; we illustrate the methods on numerous examples. A machine learning method that disregards interactions may get caught in two traps: myopia is caused by learning algorithms assuming independence in spite of interactions, whereas fragmentation arises from assuming an interaction in spite of independence.

In the standard feature selection problem, we are given a fixed set of candidate features for use in a learning problem, and must select a subset that will be used to train a model that is "as good as possible" according to some criterion. In this paper, we present an interesting and useful variant, the online feature selection problem, in which, instead of all features being available from the start, features arrive one at a time. The learner's task is to select a subset of features and return a corresponding model at each time step which is as good as possible given the features seen so far. We argue that existing feature selection methods do not perform well in this scenario, and describe a promising alternative method, based on a stagewise gradient descent technique which we call grafting.

Clustering is an essential data mining task with numerous applications. However, data in most real-life applications are high-dimensional in nature, and the related information often spreads across multiple relations. To ensure effective and efficient high-dimensional, cross-relational clustering, we propose a new approach, called CrossClus, which performs cross-relational clustering with user’s guidance. We believe that user’s guidance, even likely in very simple forms, could be essential for effective high-dimensional clustering since a user knows well the application requirements and data semantics. CrossClus is carried out as follows: a user specifies a clustering task and selects one or a small set of features pertinent to the task. CrossClus extracts the set of highly relevant features in multiple relations connected via linkages defined in the database schema, evaluates their effectiveness based on user’s guidance, and identifies interesting clusters that fit user’s needs. This method takes care of both quality in feature extraction and efficiency in clustering. Our comprehensive experiments demonstrate the effectiveness and scalability of this approach. 1.

The ability to detect and differentiate breakpoints during task execution is critical for enabling defer-to-breakpoint policies within interruption management. In this work, we examine the feasibility of building statistical models that can detect and differentiate three granularities (types) of perceptually meaningful breakpoints during task execution, without having to recognize the underlying tasks. We collected ecological samples of task execution data, and asked observers to review the interaction in the collected videos and identify any perceived breakpoints and their type. Statistical methods were applied to learn models that map features of the interaction to each type of breakpoint. Results showed that the models were able to detect and differentiate breakpoints with reasonably high accuracy across tasks. Among many uses, our resulting models can enable interruption management systems to better realize defer-to-breakpoint policies for interactive, free-form tasks.

Abstract not found