Results 1  10
of
32
Improved heterogeneous distance functions
 Journal of Artificial Intelligence Research
, 1997
"... Instancebased learning techniques typically handle continuous and linear input values well, but often do not handle nominal input attributes appropriately. The Value Difference Metric (VDM) was designed to find reasonable distance values between nominal attribute values, but it largely ignores cont ..."
Abstract

Cited by 199 (10 self)
 Add to MetaCart
Instancebased learning techniques typically handle continuous and linear input values well, but often do not handle nominal input attributes appropriately. The Value Difference Metric (VDM) was designed to find reasonable distance values between nominal attribute values, but it largely ignores continuous attributes, requiring discretization to map continuous values into nominal values. This paper proposes three new heterogeneous distance functions, called the Heterogeneous Value Difference Metric (HVDM), the Interpolated Value Difference Metric (IVDM), and the Windowed Value Difference Metric (WVDM). These new distance functions are designed to handle applications with nominal attributes, continuous attributes, or both. In experiments on 48 applications the new distance metrics achieve higher classification accuracy on average than three previous distance functions on those datasets that have both nominal and continuous attributes.
T.R.: Reduction Techniques for Instancebased Learning Algorithm
 Machine Learning
"... Abstract. Instancebased learning algorithms are often faced with the problem of deciding which instances to store for use during generalization. Storing too many instances can result in large memory requirements and slow execution speed, and can cause an oversensitivity to noise. This paper has two ..."
Abstract

Cited by 130 (2 self)
 Add to MetaCart
Abstract. Instancebased learning algorithms are often faced with the problem of deciding which instances to store for use during generalization. Storing too many instances can result in large memory requirements and slow execution speed, and can cause an oversensitivity to noise. This paper has two main purposes. First, it provides a survey of existing algorithms used to reduce storage requirements in instancebased learning algorithms and other exemplarbased algorithms. Second, it proposes six additional reduction algorithms called DROP1–DROP5 and DEL (three of which were first described in Wilson & Martinez, 1997c, as RT1–RT3) that can be used to remove instances from the concept description. These algorithms and 10 algorithms from the survey are compared on 31 classification tasks. Of those algorithms that provide substantial storage reduction, the DROP algorithms have the highest average generalization accuracy in these experiments, especially in the presence of uniform class noise. Keywords: instancebased learning, nearest neighbor, instance reduction, pruning, classification
Principal Direction Divisive Partitioning
 Data Mining and Knowledge Discovery
, 1997
"... We propose a new algorithm capable of partitioning a set of documents or other samples based on an embedding in a high dimensional Euclidean space (i.e. in which every document is a vector of real numbers). The method is unusual in that it is divisive, as opposed to agglomerative, and operates by re ..."
Abstract

Cited by 103 (22 self)
 Add to MetaCart
We propose a new algorithm capable of partitioning a set of documents or other samples based on an embedding in a high dimensional Euclidean space (i.e. in which every document is a vector of real numbers). The method is unusual in that it is divisive, as opposed to agglomerative, and operates by repeatedly splitting clusters into smaller clusters. The splits are not based on any distance or similarity measure. The documents are assembled in to a matrix which is very sparse. It is this sparsity that permits the algorithm to be very efficient. The performance of the method is illustrated with a set of text documents obtained from the World Wide Web. Some possible extensions are proposed for further investigation.
An artificial immune system architecture for computer security applications
 IEEE Transactions on Evolutionary Computation
, 2002
"... Abstract—With increased global interconnectivity, reliance on ecommerce, network services, and Internet communication, computer security has become a necessity. Organizations must protect their systems from intrusion and computervirus attacks. Such protection must detect anomalous patterns by expl ..."
Abstract

Cited by 36 (3 self)
 Add to MetaCart
Abstract—With increased global interconnectivity, reliance on ecommerce, network services, and Internet communication, computer security has become a necessity. Organizations must protect their systems from intrusion and computervirus attacks. Such protection must detect anomalous patterns by exploiting known signatures while monitoring normal computer programs and network usage for abnormalities. Current antivirus and network intrusion detection (ID) solutions can become overwhelmed by the burden of capturing and classifying new viral stains and intrusion patterns. To overcome this problem, a selfadaptive distributed agentbased defense immune system based on biological strategies is developed within a hierarchical layered architecture. A prototype interactive system is designed, implemented in Java, and tested. The results validate the use of a distributedagent biologicalsystem approach toward the computersecurity problems of virus elimination and ID. Index Terms—Agents, artificial immune system, computer security, computer virus, intrusion detection.
Location And DensityBased Hierarchical Clustering Using Similarity Analysis
 IEEE Transactions on Pattern Analysis and Machine Intelligence
, 1998
"... This paper presents a new approach to hierarchical clustering of point patterns. Two algorithms for hierarchical location and densitybased clustering are developed. Each method groups points such that maximum intracluster similarity and intercluster dissimilarity are achieved for point locations ..."
Abstract

Cited by 19 (0 self)
 Add to MetaCart
This paper presents a new approach to hierarchical clustering of point patterns. Two algorithms for hierarchical location and densitybased clustering are developed. Each method groups points such that maximum intracluster similarity and intercluster dissimilarity are achieved for point locations or point separations. Point locations or point separations are denoted as elements. Each method starts with grouping elements into clusters C e i for every element e i such that all elements in C e i are dissimilar to e i by no more than an amount `. A sample mean ¯ ¯ e i of all elements in C e i is calculated. Clusters are formed by grouping pairs of adjacent elements having similar sample means ¯ ¯ e i . Clusters with increasing intracluster similarity and intercluster dissimilarity are identified by increasing ` continuously. Those clusters that remain unchanged over intervals of ` are selected as defining a clustering of the point pattern. The accuracy and computational requirements o...
Inclusion Of Multispectral Data into Object Recognition
 Part 743 W6
, 1999
"... In this paper, we describe how object recognition benefits from exploiting multispectral and multisensor datasets. After a brief introduction we summarize the most important principles of object recognition and multisensor fusion. This serves as the basis for the proposed architecture of a multisens ..."
Abstract

Cited by 13 (4 self)
 Add to MetaCart
In this paper, we describe how object recognition benefits from exploiting multispectral and multisensor datasets. After a brief introduction we summarize the most important principles of object recognition and multisensor fusion. This serves as the basis for the proposed architecture of a multisensor object recognition system. It is characterized by multistage fusion, where the different sensory input data are processed individually and only merged at appropriate levels. The remaining sections describe the major fusion processes. Rather than providing detailed descriptions, a few examples, obtained from the Ocean City testdata site, have been chosen to illustrate the processing of the major data streams. The test site comprises of multispectral and aerial imagery, and laser scanning data. 1. INTRODUCTION The ultimate goal of digital photogrammetry is the automation of map making. This entails understanding aerial imagery and recognizing objects  both hard problems. Despite of the ...
A Micropower Programmable DSP Using Approximate Signal Processing Based on Distributed Arithmetic
 IEEE Journal of SolidState Circuits
, 2004
"... A recent trend in lowpower design has been the employment of reduced precision processing methods for decreasing arithmetic activity and average power dissipation. Such designs can trade off power and arithmetic precision as system requirements change. This work explores the potential of distribute ..."
Abstract

Cited by 13 (4 self)
 Add to MetaCart
A recent trend in lowpower design has been the employment of reduced precision processing methods for decreasing arithmetic activity and average power dissipation. Such designs can trade off power and arithmetic precision as system requirements change. This work explores the potential of distributed arithmetic (DA) computation structures for lowpower precisionondemand computation. We present an ultralowpower DSP which uses variable precision arithmetic, lowvoltage circuits, and conditional clocks to implement a biomedical detection and classification algorithm using only 560 nW. Low energy consumption enables selfpowered operation using ambient mechanical vibrations, converted to electric energy by a MEMS transducer and accompanying power electronics. The MEMS energy scavenging system is estimated to deliver 4.3 to 5.6 W of power to the DSP load.
Hierarchical Taxonomies using Divisive Partitioning
, 1998
"... We propose an unsupervised divisive partitioning algorithm for document data sets which enjoys many favorable properties. In particular, the algorithm shows excellent scalability to large data collections and produces high quality clusters which are competitive with other clustering methods. The alg ..."
Abstract

Cited by 11 (3 self)
 Add to MetaCart
We propose an unsupervised divisive partitioning algorithm for document data sets which enjoys many favorable properties. In particular, the algorithm shows excellent scalability to large data collections and produces high quality clusters which are competitive with other clustering methods. The algorithm yields information on the significant and distinctive words within each cluster, and these words can be inserted into the naturally occuring hierarchical structure produced by the algorithm. The result is an automatically generated hierarchical topical taxonomy of a document set. In this paper, we show how the algorithm's cost scales up linearly with the size of the data, illustrate experimentally the quality of the clusters produced, and show how the algorithm can produce a hierarchical topical taxonomy.
A Pattern Recognition System for Handoff Algorithms
 IEEE Journal on Selected Areas in Communication
, 2000
"... In wireless cellular systems, handoff algorithms decide when and to which base station to handoff. Traditional handoff algorithms generally cannot keep both the average number of unnecessary handoffs and the handoff decision delay low. They do not exploit the relative constancy of path loss and shad ..."
Abstract

Cited by 8 (0 self)
 Add to MetaCart
In wireless cellular systems, handoff algorithms decide when and to which base station to handoff. Traditional handoff algorithms generally cannot keep both the average number of unnecessary handoffs and the handoff decision delay low. They do not exploit the relative constancy of path loss and shadow fading effects at any given location around a base station. This information can in fact be used to improve the efficiency of handoff algorithms, as we do in our new handoff algorithms using statistical pattern recognition. Handoff algorithms with both a negligible number of unnecessary handoffs and a negligible decision delay can therefore be realized. Index TermsCellular, handoff, handover, pattern recognition, wireless. I.
Investigation of Characteristic Measures for the Analysis and Synthesis of PrecisionMachined Surfaces
 Journal of Manufacturing Systems
, 1995
"... Error prediction and control are key factors in precision machining. These factors rely on the development of formal approaches for analyzing and characterizing error sources in manufacturing. One such approach is the development of mathematical measures of precision, where precision, in this contex ..."
Abstract

Cited by 7 (7 self)
 Add to MetaCart
Error prediction and control are key factors in precision machining. These factors rely on the development of formal approaches for analyzing and characterizing error sources in manufacturing. One such approach is the development of mathematical measures of precision, where precision, in this context, is defined as surface variations of manufactured part profiles. In this paper, we discuss a novel investigation of four mathematical measures. These four methods, namely the autocorrelation function, the Fourier spectrum, a fractalwavelet representation, and the KarhunenLo`eve expansion, are applied to surfaces produced from grinding processes. The first two methods provide a basis for the investigation, as they are commonly used in the literature for qualitative signal characterization of manufacturing surfaces. However, the fractalwavelet method and KarhunenLo`eve expansion have never been applied to the analysis and synthesis of surface variations. While other fractal methods have ...