Discriminant Rule Research Articles

Impairments in set-shifting but not reversal learning in the neonatal ventral hippocampal lesion model of schizophrenia: Further evidence for medial prefrontal deficits

The executive function processes of set-shifting and reversal learning in rodents are mediated by the medial prefrontal cortex and the orbitofrontal cortex, respectively. Here, we investigated both set-shifting and reversal learning in a developmental animal model of schizophrenia, the neonatal ventral hippocampal lesion (NVHL) model. The NVHL manipulation is known to disrupt development of the medial prefrontal cortex, and to impair behaviors dependent on this area, but potential orbitofrontal abnormalities and reversal learning deficits are less well studied. Animals received excitotoxic lesions of the ventral hippocampus (NVHL) or a sham treatment during the first postnatal week, and all animals were subsequently tested in adulthood on either an operant set-shifting or an operant reversal task. Results indicated that NVHL animals were able to acquire a simple discrimination rule and exhibited normal reversal learning, but were impaired on a prefrontal-dependent set-shifting task. Furthermore, this set-shifting deficit was due to an increase in perseverative errors, which indicate difficulty suppressing a previously learned strategy and result from medial prefrontal insult. Together, these results confirm and extend the idea that cognitive impairments in the NVHL animal are primarily driven by medial prefrontal abnormalities, while the orbitofrontal cortex may remain relatively unaffected.

Finding Simple Rules for Discriminating Folding Rate Change upon Single Mutation by Statistical and Learning Methods

Protein folding rate is a valuable clue for understanding the variations in protein folding kinetics. The ability to accurately discriminate protein folding rate change is very helpful in protein design. However, there are fewer studies on the influence of amino acid substitution to protein folding rates. In our earlier studies, we constructed a dataset of 467 mutants upon amino acid substitution and proposed novel methods for discriminating and predicting the accelerating and decelerating mutants during the folding process. This study aimed to effectively develop simple rules for discriminating accelerating mutants from decelerating ones upon single amino acid substitution. The main points of the study were to build a more general dataset F661 with 661 mutants, analyze the dataset systematically, and then implement different data mining techniques to build discrimination rules. Furthermore, the rules obtained from different methods were interpreted, evaluated, compared and integrated. The results appeared that the present approach may effectively develop simple rules from these mutants and the quality of the rules may be improved by combining the statistical and learning methods. These results suggest that the present method, as well as the rules, may advance the understanding of discriminating protein folding rate change. The details of the rules along with relevant information have been integrated and available freely at http://bioinformatics.myweb.hinet.net/rulefr.htm.

Model Discrimination in Dynamic Molecular Systems: Application to Parotid De-differentiation Network

In modern systems biology the modeling of longitudinal data, such as changes in mRNA concentrations, is often of interest. Fully parametric, ordinary differential equations (ODE)-based models are typically developed for the purpose, but their lack of fit in some examples indicates that more flexible Bayesian models may be beneficial, particularly when there are relatively few data points available. However, under such sparse data scenarios it is often difficult to identify the most suitable model. The process of falsifying inappropriate candidate models is called model discrimination. We propose here a formal method of discrimination between competing Bayesian mixture-type longitudinal models that is both sensitive and sufficiently flexible to account for the complex variability of the longitudinal molecular data. The ideas from the field of Bayesian analysis of computer model validation are applied, along with modern Markov Chain Monte Carlo (MCMC) algorithms, in order to derive an appropriate Bayes discriminant rule. We restrict attention to the two-model comparison problem and present the application of the proposed rule to the mRNA data in the de-differentiation network of three mRNA concentrations in mammalian salivary glands as well as to a large synthetic dataset derived from the model used in the recent DREAM6 competition.

불균형 데이터 집합의 분류를 위한 하이브리드 SVM 모델

We call a data set in which the number of records belonging to a certain class far outnumbers the number of records belonging to the other class, `imbalanced data set`. Most of the classification techniques perform poorly on imbalanced data sets. When we evaluate the performance of a certain classification technique, we need to measure not only `accuracy` but also `sensitivity` and `specificity`. In a customer churn prediction problem, `retention` records account for the majority class, and `churn` records account for the minority class. Sensitivity measures the proportion of actual retentions which are correctly identified as such. Specificity measures the proportion of churns which are correctly identified as such. The poor performance of the classification techniques on imbalanced data sets is due to the low value of specificity. Many previous researches on imbalanced data sets employed `oversampling` technique where members of the minority class are sampled more than those of the majority class in order to make a relatively balanced data set. When a classification model is constructed using this oversampled balanced data set, specificity can be improved but sensitivity will be decreased. In this research, we developed a hybrid model of support vector machine (SVM), artificial neural network (ANN) and decision tree, that improves specificity while maintaining sensitivity. We named this hybrid model `hybrid SVM model.` The process of construction and prediction of our hybrid SVM model is as follows. By oversampling from the original imbalanced data set, a balanced data set is prepared. SVM_I model and ANN_I model are constructed using the imbalanced data set, and SVM_B model is constructed using the balanced data set. SVM_I model is superior in sensitivity and SVM_B model is superior in specificity. For a record on which both SVM_I model and SVM_B model make the same prediction, that prediction becomes the final solution. If they make different prediction, the final solution is determined by the discrimination rules obtained by ANN and decision tree. For a record on which SVM_I model and SVM_B model make different predictions, a decision tree model is constructed using ANN_I output value as input and actual retention or churn as target. We obtained the following two discrimination rules: `IF ANN_I output value

License Plate Automatic Recognition Model based on PCA-SVM

License plate recognition is one of the important research topic in computer vision, image processing and pattern recognition technology in the field of intelligent transportation applications. It is an important part of the intelligent traffic management, project management of highways, urban transport and parking and plays a decisive role. In order to increase the license plate recognition precisions, this paper proposes a license plate automatic recognition model (PCA-SVM) which combines principal component analysis and Support Vector Machine. First, the structural features and image grey features, such as license plate outline and strokes orders are extracted; then, the principal component analysis fuses the two types of features, reduces the dimensions and removes redundancy. Finally, Support Vector Machine is applied to establish structural feature recognition model and grey feature recognition model and final recognition results are evaluated by high confidence discrimination rules. The simulation results illustrate PCA-SVM can increase the license plate recognition precisions, and the recognition speed is fast. It can be applied in practical license plate recognition.

Performance and estimation of the true error rate of classification rules built with additional information. An application to a cancer trial

Classification rules that incorporate additional information usually present in discrimination problems are receiving certain attention during the last years as they perform better than the usual rules. Fernández, M. A., C. Rueda and B. Salvador (2006): "Incorporating additional information to normal linear discriminant rules," J. Am. Stat. Assoc., 101, 569-577, proved that these rules have lower total misclassification probability than the usual Fisher's rule. In this paper we consider two issues; on the one hand, we compare these rules with those based on shrinkage estimators of the mean proposed by Tong, T., L. Chen and H. Zhao (2012): "Improved mean estimation and its application to diagonal discriminant analysis," Bioinformatics, 28(4): 531-537. with regard to four criteria: total misclassification probability, area under ROC curve, well-calibratedness and refinement; on the other hand, we consider the estimation of the true error rate, which is a very interesting parameter in applications. We prove results on the apparent error rate of the rules that expose the need of new estimators of their true error rate. We propose four such new estimators. Two of them are defined incorporating the additional information into the leave-one-out-bootstrap. The other two are the corresponding cross-validation after bootstrap versions. We compare these estimators with the usual ones in a simulation study and in a cancer trial application, showing the good behavior of the rules that incorporate additional information and of the new leave-one-out bootstrap estimators of their true error rate.

Dissociable roles of the dorsal striatum and dorsal hippocampus in conditional discrimination and spatial alternation T-maze tasks

The roles of the dorsal hippocampus (DH) and dorsal striatum (DS) in the learning and retention of conditional discrimination (CD) rules is a subject of debate. Although previous studies have examined the relationship between the DH and DS and the performance of CD tasks in operant chambers, the relative contributions of these two brain regions to the retention of CD rules requiring an association between a cue and a spatial location have not been characterized. We designed an experiment to assess the roles of the DH and DS in the retention of a visuospatial CD task by transiently inactivating either structure with muscimol in separate groups of rats and measuring performance on a previously learned CD task. The performance of two other groups of rats on a previously learned delayed spatial alternation (DA) task was also measured following inactivation of either DS or DH, which allowed us to control for any possibly confounding effects of spatial cues present in the testing room, length of the intertrial interval period on the performance of the CD task, and muscimol on sensorimotor or motivational processing. Muscimol inactivation of dorsal striatum, but not dorsal hippocampus, impaired CD performance, while inactivation of dorsal hippocampus, but not dorsal striatum impaired DA performance. These results demonstrate a double dissociation between the roles of the DH and DS in these two tasks, and provide a systematic characterization of the relationship between these two brain areas and CD performance.

Relational Learning in Glaucous-Winged Gulls (Larus glaucescens)

An experimental approach was created for the comparative investigation of the cognitive abilities of the glaucous-winged gull (Larus glaucescens) in their natural habitat. The territoriality of gulls during the breeding period and the fact that the gulls inhabiting the territory of the Komandorsky Reserve are practically not in fear of humans allowed us to work with individually recognized birds directly at their nest sites inside the colony. The possibility of using this approach to investigate their cognitive abilities was demonstrated on 24 gulls, in particular, to investigate their abilities for relative size generalization. The first experiment illustrated that the gulls are able to learn to discriminate two pairs of stimuli according to the feature: 'larger' or 'smaller'. They were then given a test to transfer the discriminative rule in which novel combinations of the same stimuli were used. The gulls successfully coped with only a few of these tests. In the next experiment the birds were taught to discriminate four pairs of similar stimuli. The majority of the birds coped with the tests to transfer the discriminative rule both to the novel combinations of familiar stimuli, and also to the novel stimuli of the familiar category (items of different colour and shape). However, none of the birds transferred the discriminative rule to stimuli of a novel category (sets differing by number of components). Thus, in their ability to generalize at a preconceptual level gulls are more comparable with pigeons, whereas large-brained birds (crows and parrots), are capable of concept formation.

Energy discrimination and international rules in hard times: what's new this time around, and what can be done

International trade and investment in the energy sector takes place in a framework of treaties based on the conviction that all boats rise on the tide of free trade and underpinned by legal obligations, including the commitment to treat foreign products no less favourably than domestic products. Although international treaty obligations should limit violations of basic trade rules, in times of economic crisis domestic political imperatives of collecting revenue and protecting local employment often trump economic wisdom and lead to violations of international trade treaties. That the current global economic crisis has produced its share of trade protectionism comes as no surprise. However, the ever expanding substantive and geographic scope of international trade rules forces governments to seek out increasingly nuanced and creative modes of protection in order to disguise treaty violations. The recent trend toward unapologetic protectionism may indicate a new direction in trade policy. This article reviews trade protectionism that affects businesses in the energy sector, discusses attendant violations of international trade law, and identifies actions that could be taken to counter trade protectionism affecting the energy sector.

Generalization Error of Linear Discriminant Analysis in Spatially-Correlated Sensor Networks

Generalization error, the probability of error of a detection rule learned from training samples on new unseen samples, is a fundamental quantity to be characterized. However, characterizations of generalization error in the statistical learning theory literature are often loose and practically unusable for optimizing detection systems. In this work, focusing on learning linear discriminant analysis detection rules from spatially-correlated sensor measurements, a tight generalization error approximation is developed that can be used to optimize the parameters of a sensor network detection system. As such, the approximation is used to optimize network settings. The approximation is also used to derive a detection error exponent and select an optimal subset of deployed sensor nodes. A Gauss-Markov random field is used to model correlation and weak laws of large numbers in geometric probability are employed in the analysis.

A Road to Classification in High Dimensional Space: The Regularized Optimal Affine Discriminant

SummaryFor high dimensional classification, it is well known that naively performing the Fisher discriminant rule leads to poor results due to diverging spectra and accumulation of noise. Therefore, researchers proposed independence rules to circumvent the diverging spectra, and sparse independence rules to mitigate the issue of accumulation of noise. However, in biological applications, often a group of correlated genes are responsible for clinical outcomes, and the use of the covariance information can significantly reduce misclassification rates. In theory the extent of such error rate reductions is unveiled by comparing the misclassification rates of the Fisher discriminant rule and the independence rule. To materialize the gain on the basis of finite samples, a regularized optimal affine discriminant (ROAD) is proposed. The ROAD selects an increasing number of features as the regularization relaxes. Further benefits can be achieved when a screening method is employed to narrow the feature pool before applying the ROAD method. An efficient constrained co-ordinate descent algorithm is also developed to solve the associated optimization problems. Sampling properties of oracle type are established. Simulation studies and real data analysis support our theoretical results and demonstrate the advantages of the new classification procedure under a variety of correlation structures. A delicate result on continuous piecewise linear solution paths for the ROAD optimization problem at the population level justifies the linear interpolation of the constrained co-ordinate descent algorithm.

Development and application of an ecological classification tool for fish in lakes in Ireland

A classification tool suitable for establishing the ecological status of lakes based on fish population parameters has been developed for the Republic of Ireland and Northern Ireland (EU Water Framework Directive Ecoregion 17). A lake typology relevant to fish populations in lakes from Ecoregion 17 was produced as part of the ecological classification tool development. Four lake types were determined based on fish metrics and abiotic variables from 43 “reference” lakes. The specific lake fish typology categorised lakes into low (≤67CaCO3mgL−1) or high (>67CaCO3mgL−1) alkalinity, and shallow (≤17m) or deep (>17m) maximum depth. The fish in lakes classification tool (FIL2) follows a novel multimetric predictive approach, assigning ecological status to a lake using two independent methods. FIL2 qualitatively defines a lake's ecological status based on fish metrics using discriminant classification rules and, using a generalised linear model, quantitatively derives an Ecological Quality Ratio (EQR, 0<EQR<1), along with associated confidence intervals. It is recommended that both methods are used to validate output and cross-check and highlight potential misclassification.

Rapid and non-destructive discrimination of tea varieties by near infrared diffuse reflection spectroscopy coupled with classification and regression trees

The current study attempted to rapidly and non-destructively discriminate the diverse varieties of tea (that is, Biluochun, Longjing, Maojian, Qihong, Tieguanyin, and Yinzhen) via utilizing near infrared (NIR) diffuse reflectance spectroscopy coupled with pattern recognition strategies. Before the recognition analysis, the original NIR spectra were pre-processed by second derivative treatment followed by informative wavenumber interval location. And then, non-linearity detection and outlier diagnosis were performed. When pattern recognition referred, principal component analysis (PCA) was firstly applied to ascertain the discrimination possibility with the NIR spectra. Classification and regression trees (CART), compared with linear discriminant analysis (LDA), and partial squares-discriminant analysis (PLS-DA), was then employed for establishing the discrimination rule. Experimental results showed that the tea quality could be accurately, rapidly, and non-invasively identified via NIR spectroscopy coupled with CART. Key words : Near infrared diffuse reflection spectroscopy, classification and regression trees, and tea variety discrimination.

Rigid-Motion-Invariant Classification of 3-D Textures

This paper studies the problem of 3-D rigid-motion-invariant texture discrimination for discrete 3-D textures that are spatially homogeneous by modeling them as stationary Gaussian random fields. The latter property and our formulation of a 3-D rigid motion of a texture reduce the problem to the study of 3-D rotations of discrete textures. We formally develop the concept of 3-D texture rotations in the 3-D digital domain. We use this novel concept to define a "distance" between 3-D textures that remains invariant under all 3-D rigid motions of the texture. This concept of "distance" can be used for a monoscale or a multiscale 3-D rigid-motion-invariant testing of the statistical similarity of the 3-D textures. To compute the "distance" between any two rotations R(1) and R(2) of two given 3-D textures, we use the Kullback-Leibler divergence between 3-D Gaussian Markov random fields fitted to the rotated texture data. Then, the 3-D rigid-motion-invariant texture distance is the integral average, with respect to the Haar measure of the group SO(3), of all of these divergences when rotations R(1) and R(2) vary throughout SO(3). We also present an algorithm enabling the computation of the proposed 3-D rigid-motion-invariant texture distance as well as rules for 3-D rigid-motion-invariant texture discrimination/classification and experimental results demonstrating the capabilities of the proposed 3-D rigid-motion texture discrimination rules when applied in a multiscale setting, even on very general 3-D texture models.

A study of the effect of kurtosis on discriminant analysis under elliptical populations

This paper is concerned with the role some parameters indexing four important families within the multivariate elliptically contoured distributions play as indicators of multivariate kurtosis. The problem is addressed for the exponential power family, for a subclass of the Kotz family and for the Pearson type II and type VII distributions. Once such a problem is analyzed, we study the effect these parameters have, as kurtosis indicators, on binary discriminant analysis by exploring their relationship with the error rate of the Bayes discriminant rule. The effect is analyzed under mild conditions on the kernel function generating the elliptical density. Some numerical examples are given in order to illustrate our theoretical insights and findings.

The Law and Economics of Network Neutrality

The Federal Communications Commission's Network Neutrality Order regulates how broadband networks explain their services to customers, mandates that subscribers be permitted to deploy whatever computers, mobile devices, or applications they like for use with the network access service they purchase, imposes a prohibition upon unreasonable discrimination in network management such that Internet Service Provider efforts to maintain service quality (e.g. mitigation congestion) or to price and package their services do not burden rival applications. This paper offers legal and economic critique of the new Network Neutrality policy and particularly the no blocking and no discrimination rules. While we argue the FCC‘s rules are likely to be declared beyond the scope of the agency's charter, we focus upon the economic impact of net neutrality regulations. It is beyond paradoxical that the FCC argues that it is imposing new regulations so as to preserve the Internet's current economic structure; that structure has developed in an unregulated environment where firms are free to experiment with business models - and vertical integration - at will. We demonstrate that Network Neutrality goes far further than existing law, categorically prohibiting various forms of economic integration in a manner equivalent to antitrust's per se rule, properly reserved for conduct that is so likely to cause competitive harm that the marginal benefit of a fact-intensive analysis cannot be justified. Economic analysis demonstrates that Network Neutrality cannot be justified upon consumer welfare grounds. Further, the Commission's attempt to justify its new policy simply ignores compelling evidence that “open access” regulations have distorted broadband build-out in the United States, visibly reducing subscriber growth when imposed and visibly increasing subscriber growth when repealed. On the other, the FCC manages to cite just one study - not of the broadband market - to support its claims of widespread foreclosure threats. This empirical study, upon closer scrutiny than the Commission appears to have given it, actually shows no evidence of anti-competitive foreclosure. This fatal analytical flaw constitutes a smoking gun in the FCC's economic analysis of net neutrality.

Constrained linear discriminant rule via the Studentized classification statistic based on monotone missing data

This paper provides an asymptotic expansion for the distribution of the Studentized linear discriminant function with k-step monotone missing training data. It turns out to be a certain generalization of the results derived by Anderson [1] and Shutoh and Seo [12]. Furthermore we also derive the cut-off point constrained by a conditional probability of misclassification using the idea of McLachlan [8]. Finally we perform Monte Carlo simulation to evaluate our results.

Improved mean estimation and its application to diagonal discriminant analysis

High-dimensional data such as microarrays have created new challenges to traditional statistical methods. One such example is on class prediction with high-dimension, low-sample size data. Due to the small sample size, the sample mean estimates are usually unreliable. As a consequence, the performance of the class prediction methods using the sample mean may also be unsatisfactory. To obtain more accurate estimation of parameters some statistical methods, such as regularizations through shrinkage, are often desired. In this article, we investigate the family of shrinkage estimators for the mean value under the quadratic loss function. The optimal shrinkage parameter is proposed under the scenario when the sample size is fixed and the dimension is large. We then construct a shrinkage-based diagonal discriminant rule by replacing the sample mean by the proposed shrinkage mean. Finally, we demonstrate via simulation studies and real data analysis that the proposed shrinkage-based rule outperforms its original competitor in a wide range of settings.

Late Talkers: Do Good Predictors of Outcome Exist?

Both small-scale and epidemiological longitudinal studies of early language delay indicate that most late talkers attain language scores in the average range by age 5, 6, or 7. However, late talker groups typically obtain significantly lower scores than groups with typical language histories on most language measures into adolescence. These findings support a dimensional account of language delay, whereby late talkers and typically developing peers differ quantitatively on a hypothetical language ability spectrum. Variation in language ability is presumed to derive from variation in skills subserving language, such as auditory perception/processing, word retrieval, verbal working memory, motor planning, phonological discrimination, and grammatical rule learning. Expressive language screening at 18-35 months can serve an important public health function by identifying children whose expressive delay is secondary to autism spectrum disorder, intellectual disability, hearing impairment, receptive language delay, or demographic risk. Finally, the review suggests that demographic risk associated with low SES may become more important as a causal factor in language delay as children get older.

The Law and Economics of Network Neutrality

The Federal Communications Commission's Network Neutrality Order regulates how broadband networks explain their services to customers, mandates that subscribers be permitted to deploy whatever computers, mobile devices, or applications they like for use with the network access service they purchase, imposes a prohibition upon unreasonable discrimination in network management such that Internet Service Provider efforts to maintain service quality (e.g. mitigation congestion) or to price and package their services do not burden rival applications. This paper offers legal and economic critique of the new Network Neutrality policy and particularly the no blocking and no discrimination rules. While we argue the FCC‘s rules are likely to be declared beyond the scope of the agency's charter, we focus upon the economic impact of net neutrality regulations. It is beyond paradoxical that the FCC argues that it is imposing new regulations so as to preserve the Internet's current economic structure; that structure has developed in an unregulated environment where firms are free to experiment with business models - and vertical integration - at will. We demonstrate that Network Neutrality goes far further than existing law, categorically prohibiting various forms of economic integration in a manner equivalent to antitrust's per se rule, properly reserved for conduct that is so likely to cause competitive harm that the marginal benefit of a fact-intensive analysis cannot be justified. Economic analysis demonstrates that Network Neutrality cannot be justified upon consumer welfare grounds. Further, the Commission's attempt to justify its new policy simply ignores compelling evidence that “open access” regulations have distorted broadband build-out in the United States, visibly reducing subscriber growth when imposed and visibly increasing subscriber growth when repealed. On the other, the FCC manages to cite just one study - not of the broadband market - to support its claims of widespread foreclosure threats. This empirical study, upon closer scrutiny than the Commission appears to have given it, actually shows no evidence of anti-competitive foreclosure. This fatal analytical flaw constitutes a smoking gun in the FCC's economic analysis of net neutrality.