Accurate ensemble pruning with PL-bagging

Ensemble pruning deals with the selection of base learners prior to combination in order to improve prediction accuracy and efficiency. In the ensemble literature, it has been pointed out that in order for an ensemble classifier to achieve higher prediction accuracy, it is critical for the ensemble classifier to consist of accurate classifiers which at the same time diverse as much as possible. In this paper, a novel ensemble pruning method, called PL-bagging, is proposed. In order to attain the balance between diversity and accuracy of base learners, PL-bagging employs positive Lasso to assign weights to base learners in the combination step. Simulation studies and theoretical investigation showed that PL-bagging filters out redundant base learners while it assigns higher weights to more accurate base learners. Such improved weighting scheme of PL-bagging further results in higher classification accuracy and the improvement becomes even more significant as the ensemble size increases. The performance of PL-bagging was compared with state-of-the-art ensemble pruning methods for aggregation of bootstrapped base learners using 22 real and 4 synthetic datasets. The results indicate that PL-bagging significantly outperforms state-of-the-art ensemble pruning methods such as Boosting-based pruning and Trimmed bagging.     

Stream Classification System Based on Susceptibility to Algal Growth in Response to Nutrients

We developed a stream classification system that is based on stream’s susceptibility to algal growth using a two-step approach. The model portrays algal biomass as a result of stream’s response to nutrient concentrations and the response is governed by various stream factors. In the first step, a nutrient-chlorophyll a relationship was developed to characterize nutrient’s effects on algal biomass. Residuals of the relationship were attributed to stream’s susceptibility to algal growth in response to nutrients and referred to as “observed” susceptibility. In the second step, conditions of other contributing factors were used to explain the variation in the residuals and the developed relationship was used to generate “predicted” susceptibility. Existing data compiled from various monitoring projects of Illinois streams and rivers were used to illustrate the approach. Streams were classified into three (high, medium, and low) categories based on their observed and predicted susceptibility values, respectively. With the available data, the model showed a 40-50% success rate for classifying the streams based on three observed and predicted susceptibility categories. Model entropy also was calculated for selecting the best model. The results show the important role of both nutrients and other contributing factors in explaining the variation of algal biomass. The study also suggests ways to fine tune the model and improve its accuracy, which would make the presented model a more viable tool for stream classification for establishing nutrient criteria to prevent surface streams from eutrophication.     

Generation and selection of likely interpretations during plan recognition in task-oriented consultation systems

This paper presents a mechanism which infers a user's plans from his/her utterances by directing the inference process towards the more likely interpretations of a speaker's statements among many possible interpretations. Our mechanism uses Bayesian theory of probability to assess the likelihood of an interpretation, and it complements this assessment by taking into consideration two aspects of an interpretation: its coherence and its information content. The coherence of an interpretation is determined by the relationships between the different statements in the discourse. The information content of an interpretation is a measure of how well defined the interpretation is in terms of the actions to be performed on the basis of this interpretation. This measure is used to guide the inference process towards interpretations with higher information content. The information content of an interpretation depends on the specificity and the certainty of the inferences in it, where the certainty of an inference depends on the knowledge on which the inference is based. Our mechanism has been developed for use in task-oriented consultation systems. The particular domain that we have chosen for exploration is that of travel booking.     

Data mining of resilience indicators

In recent years, the Asia-Pacific region has experienced several financial setbacks, including speculative attacks in 1998 and the SARS outbreak in 2003. Financial stresses of this nature are unanticipated, and not all of the dangers can be predicted by the examination of market information and macroeconomic indicators. The Early Warning System (EWS) that has been adopted by the International Monetary Fund may not be able to predict future financial crises for all possible scenarios, because shocks come in many different forms. To supplement the EWS, this paper proposes a data mining framework to measure the resilience of an economy. The resilience framework does not predict a crisis, but rather assesses the current state of health of an economy and its ability to withstand a financial shock should one occur. The framework is based on a feedback system consisting of two stages. The first stage assigns a resilience score to each economy based on a fuzzy logic scoring scheme that is built on the ambiguous reasoning of experts. The second stage uses the classification tree approach to estimate thresholds for each economic indicator, and examines the quality of the fuzzy score. The result from the second stage is then passed back to the first stage as feedback. The final result is obtained when the feedback system reaches its equilibrium state. The proposed resilience framework is applied to the external-sector and the public-sector economies of several countries to illustrate its applicability.     

一种协调的科技文献分类方法

科技文献之间的相互引证关系反映了一种科学交流活动,显示了科学文献之间(甚至是学科之间)的内在联系,而通过追溯文献之间的这种关系,可以改善和提高传统的基于内容的科技文献的分类的方法。论文利用有相互引证关系,有同引关系,以及有耦合关系的两篇文献一般是属于同一类的这一特点,提出了文献之间的引用相似度,同引相似度,耦合相似度这三个概念,再利用这三个概念生成了文献之间的“结构相似度”,并将它用于K-NN分类法中得出一种基于结构的分类法。最后,论文将这种基于结构的分类法和基于内容的NaveBayes分类法结合起来提出了一种新的协调分类法。     

Experiments of Fast Learning with High Order Boltzmann Machines

This work reports the results obtained with the application of High Order Boltzmann Machines without hidden units to construct classifiers for some problems that represent different learning paradigms. The Boltzmann Machine weight updating algorithm remains the same even when some of the units can take values in a discrete set or in a continuous interval. The absence of hidden units and the restriction to classification problems allows for the estimation of the connection statistics, without the computational cost involved in the application of simulated annealing. In this setting, the learning process can be sped up several orders of magnitude with no appreciable loss of quality of the results obtained.     

人类HLA-DRB1基因型与民族和人群的神经网络无监督模式分类

HLA-DRB1等位基因主要参与人类的抗原免疫功能,对其进行比较研究,有助于追溯人类的进化迁移史,开发类群特异性药物.本研究以Matlab为平台,用自组织竞争网络(Self-Organizing Competitive Neural Network),对世界54个民族和人群、14个HLA-DRB1等位基因,进行了无监督模式分类.结果表明,各民族之间存在差异性,同民族的各人群之间有相似性.西伯利亚各人群,澳洲各土著人群,黑人各人群,南美印第安各人群,犹太族各人群,日本各人群,及欧美白人各人群有相对独立性:南美印第安人与西伯利亚人有高度相似性;中国民族在南方人群和北方人群间存在较大差异;中国汉族中,广东汉族地位特殊,与少数民族如拉祜族、瑶族关系密切.     

Motor oil classification by base stock and viscosity based on near infrared (NIR) spectroscopy data

In this paper we have tried to build effective model for classification of motor oils by base stock and viscosity class. Three (3) sets of near infrared (NIR) spectra (1125, 1010, and 1050 spectra) were used for classification of motor oils into 3 or 4 classes according to their base stock (synthetic, semi-synthetic, and mineral), kinematic viscosity at low temperature (SAE 0W, 5W, 10W, and 15W) and kinematic viscosity at high temperature (SAE 20, 30, 40, and 50). The abilities of three (3) different classification methods: regularized discriminant analysis (RDA), soft independent modelling of class analogy (SIMCA), and multilayer perceptron (MLP) - were also compared. In all cases NIR spectroscopy was found to be quite effective for motor oil classification. MLP classification technique was found to be the most effective one.     

Linear mixture model applied to Amazonian vegetation classification

Many research projects require accurate delineation of different secondary succession (SS) stages over large regions/subregions of the Amazon basin. However, the complexity of vegetation stand structure, abundant vegetation species, and the smooth transition between different SS stages make vegetation classification difficult when using traditional approaches such as the maximum likelihood classifier (MLC). Most of the time, classification distinguishes only between forest/non-forest. It has been difficult to accurately distinguish stages of SS. In this paper, a linear mixture model (LMM) approach is applied to classify successional and mature forests using Thematic Mapper (TM) imagery in the Rondônia region of the Brazilian Amazon. Three endmembers (i.e., shade, soil, and green vegetation or GV) were identified based on the image itself and a constrained least-squares solution was used to unmix the image. This study indicates that the LMM approach is a promising method for distinguishing successional and mature forests in the Amazon basin using TM data. It improved vegetation classification accuracy over that of the MLC. Initial, intermediate, and advanced successional and mature forests were classified with overall accuracy of 78.2% using a threshold method on the ratio of shade to GV fractions, a 7.4% increase over the MLC. The GV and shade fractions are sensitive to the change of vegetation stand structures and better capture biophysical structure information.