Pattern Recognition and Machine Learning
TL;DR: This book covers a broad range of topics for regular factorial designs and presents all of the material in very mathematical fashion and will surely become an invaluable resource for researchers and graduate students doing research in the design of factorial experiments.
Abstract: (2007). Pattern Recognition and Machine Learning. Technometrics: Vol. 49, No. 3, pp. 366-366.
Citations
More filters
01 Nov 2011
TL;DR: Kobe is a tool that aids mobile classifier development with the help of a SQL-like programming interface that enables easier development of mobile sensing and classification apps.
Abstract: Sensor convergence on the mobile phone is spawning a broad base of new and interesting mobile applications. As applications grow in sophistication, raw sensor readings often require classification into more useful application-specific high-level data. For example, GPS readings can be classified as running, walking or biking. Unfortunately, traditional classifiers are not built for the challenges of mobile systems: energy, latency, and the dynamics of mobile.Kobe is a tool that aids mobile classifier development. With the help of a SQL-like programming interface, Kobe performs profiling and optimization of classifiers to achieve an optimal energy-latency-accuracy tradeoff. We show through experimentation on five real scenarios, classifiers on Kobe exhibit tight utilization of available resources. For comparable levels of accuracy traditional classifiers, which do not account for resources, suffer between 66% and 176% longer latencies and use between 31% and 330% more energy. From the experience of using Kobe to prototype two new applications, we observe that Kobe enables easier development of mobile sensing and classification apps.
119 citations
Cites background from "Pattern Recognition and Machine Lea..."
...To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior speci.c permission and/or a fee....
[...]
TL;DR: In this article, a new framework for supervisory protection and situational awareness to enhance grid operations and protection using modern wide-area monitoring systems is presented, which analyzes only the PMU data with the strongest or the most prominent disturbance signature.
Abstract: This paper presents a new framework for supervisory protection and situational awareness to enhance grid operations and protection using modern wide-area monitoring systems. In contrast to earlier approaches dealing with the combined processing of data from multiple phasor measurement units (PMUs), the proposed approach analyzes only the PMU data with the strongest or the most prominent disturbance signature. The specific contributions of this paper are: a) new criteria for identification of PMU with the strongest signature, b) simplified approach for quick detection of faults, c) early classification of eight other disturbances suitable for near real-time response, d) time-frequency transform-based feature extraction techniques for speedy and reliable classifiers, and e) a promising approach to locate disturbances within narrow geographical constraints. The contributions are verified with exhaustive simulation data from the Western Electricity Coordination Council system model and limited real PMU data.
119 citations
Cites background from "Pattern Recognition and Machine Lea..."
...The number of hidden layer nodes is a parameter that can be adjusted to give the best predictive performance [20]....
[...]
01 Dec 2011
TL;DR: It is demonstrated that existing approaches to the use of EVT for novelty detection are appropriate only for univariate, unimodal problems and a principled approach to the analysis of high-dimensional data to be taken.
Abstract: Novelty detection, or one-class classification, aims to determine if data are "normal" with respect to some model of normality constructed using examples of normal system behaviour. If that model is composed of generative probability distributions, the extent of "normality" in the data space can be described using Extreme Value Theory (EVT), a branch of statistics concerned with describing the tails of distributions. This paper demonstrates that existing approaches to the use of EVT for novelty detection are appropriate only for univariate, unimodal problems. We generalise the use of EVT for novelty detection to the analysis of data with multivariate, multimodal distributions, allowing a principled approach to the analysis of high-dimensional data to be taken. Examples are provided using vital-sign data obtained from a large clinical study of patients in a high-dependency hospital ward.
119 citations
Cites methods from "Pattern Recognition and Machine Lea..."
...Typically, the underlying distribution fn is multivariate and multimodal; it could be, for example, approximated using a Gaussian mixture model (GMM), a Parzen window estimator, or some other mixture of components [2]....
[...]
TL;DR: In this article, a systematic search for extremely metal-poor (XMP) galaxies in the spectroscopic sample of SDSS data release 7 (DR7) was carried out by classifying all the galaxies according to the form of their spectra.
Abstract: We carry out a systematic search for extremely metal-poor (XMP) galaxies in the spectroscopic sample of Sloan Digital Sky Survey (SDSS) data release 7 (DR7). The XMP candidates are found by classifying all the galaxies according to the form of their spectra in a region 80 ? wide around H?. Due to the data size, the method requires an automatic classification algorithm. We use k-means. Our systematic search renders 32 galaxies having negligible [N II] lines, as expected in XMP galaxy spectra. Twenty-one of them have been previously identified as XMP galaxies in the literature?the remaining 11 are new. This was established after a thorough bibliographic search that yielded only some 130 galaxies known to have an oxygen metallicity 10 times smaller than the Sun (explicitly, with 12 + log (O/H) ? 7.65). XMP galaxies are rare; they represent 0.01% of the galaxies with emission lines in SDSS/DR7. Although the final metallicity estimate of all candidates remains pending, strong-line empirical calibrations indicate a metallicity about one-tenth solar, with the oxygen metallicity of the 21 known targets being 12 + log (O/H) 7.61 ? 0.19. Since the SDSS catalog is limited in apparent magnitude, we have been able to estimate the volume number density of XMP galaxies in the local universe, which turns out to be (1.32 ? 0.23) ? 10?4 Mpc?3. The XMP galaxies constitute 0.1% of the galaxies in the local volume, or ~0.2% considering only emission-line galaxies. All but four of our candidates are blue compact dwarf galaxies, and 24 of them have either cometary shape or are formed by chained knots.
119 citations
Cites methods from "Pattern Recognition and Machine Lea..."
...It is a rather standard technique in data-mining, machine learning, and artificial intelligence (e.g., Everitt 1995; Bishop 2006), and we have already successfully employed it for massive classification of galaxy spectra (Sánchez Almeida et al....
[...]
...It is a rather standard technique in data-mining, machine learning, and artificial intelligence (e.g., Everitt 1995; Bishop 2006), and we have already successfully employed it for massive classification of galaxy spectra (Sánchez Almeida et al. 2008, 2010)....
[...]
TL;DR: A new EEG feature selection approach, transfer recursive feature elimination (T-RFE), to determine a set of the most robust EEG indicators with stable geometrical distribution across a group of training subjects and a specific testing subject and the effectiveness of the T-R FE algorithm for such cross-subject emotion classification paradigm has been validated by DEAP database.
Abstract: Using machine-learning methodologies to analyze EEG signals becomes increasingly attractive for recognizing human emotions because of the objectivity of physiological data and the capability of the learning principles on modeling emotion classifiers from heterogeneous features. However, the conventional subject-specific classifiers may induce additional burdens to each subject for preparing multiple-session EEG data as training sets. To this end, we developed a new EEG feature selection approach, transfer recursive feature elimination (T-RFE), to determine a set of the most robust EEG indicators with stable geometrical distribution across a group of training subjects and a specific testing subject. A validating set is introduced to independently determine the optimal hyper-parameter and the feature ranking of the T-RFE model aiming at controlling the overfitting. The effectiveness of the T-RFE algorithm for such cross-subject emotion classification paradigm has been validated by DEAP database. With a linear least square support vector machine classifier implemented, the performance of the T-RFE is compared against several conventional feature selection schemes and the statistical significant improvement has been found. The classification rate and F-score achieve 0.7867, 0.7526, 0.7875, and 0.8077 for arousal and valence dimensions, respectively, and outperform several recent reported works on the same database. In the end, the T-RFE based classifier is compared against two subject-generic classifiers in the literature. The investigation of the computational time for all classifiers indicates the accuracy improvement of the T-RFE is at the cost of the longer training time.
119 citations
Cites background from "Pattern Recognition and Machine Lea..."
...Regarding to the fact that the cross-subject emotion recognition is a typical domain adaptation problem in transfer learning (Bishop, 2006), it is natural to generalize the conventional RFE to the T-RFE aiming at transferring common knowledge across two or more different subjects in a shared lowdimensional feature space....
[...]