Filters








704,126 Hits in 5.0 sec

Variable selection in model-based discriminant analysis

C. Maugis, G. Celeux, M.-L. Martin-Magniette
2011 Journal of Multivariate Analysis  
In particular, it is shown that this well ground variable selection model can be of great interest to improve the classification performance of the quadratic discriminant analysis in a high dimension context  ...  This variable selection model was inspired by a previous work on variable selection in model-based clustering. A BIC-like model selection criterion is proposed.  ...  Considering the classification problem in the model-based discriminant analysis context allows us to recast variable selection into a model selection problem and to adapt the variable selection model for  ... 
doi:10.1016/j.jmva.2011.05.004 fatcat:k47dj3usojhvhk47ssfnftj6va

Model-based mixture discriminant analysis—an experimental study

Zohar Halbe, Mayer Aladjem
2005 Pattern Recognition  
The subject of this paper is an experimental study of a discriminant analysis (DA) based on Gaussian mixture estimation of the class-conditional densities.  ...  Introduction Discriminant analysis (DA) is a powerful technique for classifying observations into known preexisting classes.  ...  Model-based DA Following [2, 3] Table 1 gives the expressions for ν c , for the parameterizations of Σ cj used in this paper.  ... 
doi:10.1016/j.patcog.2004.08.010 fatcat:544d4qhp4faipnhq5kiiw5ql3e

Discriminative Models Can Still Outperform Generative Models in Aspect Based Sentiment Analysis [article]

Dhruv Mullick, Alona Fyshe, Bilal Ghanem
2022 arXiv   pre-print
Aspect-based Sentiment Analysis (ABSA) helps to explain customers' opinions towards products and services.  ...  Previous results showed that generative models outperform discriminative models on several English ABSA datasets.  ...  A.4 Error Analysis We conduct an error analysis on the outputs of the models to better understand the cases where they fail.  ... 
arXiv:2206.02892v1 fatcat:kev7gac4w5drbjimwvczlf2jc4

Fluid Dynamic Models for Bhattacharyya-Based Discriminant Analysis

Yung-Kyun Noh, Jihun Hamm, Frank Chongwoo Park, Byoung-Tak Zhang, Daniel D. Lee
2018 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Here, we investigate a physics-based model where we consider the labeled data as interacting fluid distributions.  ...  Unfortunately, the optimal approach of finding the low-dimensional projection with minimal Bayes classification error is intractable, so most standard algorithms optimize a tractable heuristic function  ...  We show that our physical fluid discriminant model approximates the optimal Bayes error criterion in both of these special cases. The remainder of the paper is organized as follows.  ... 
doi:10.1109/tpami.2017.2666148 pmid:28186879 fatcat:ifurm7zlbbe7jblnna4bsvduty

Object Detection with Discriminatively Trained Part-Based Models

P F Felzenszwalb, R B Girshick, D McAllester, D Ramanan
2010 IEEE Transactions on Pattern Analysis and Machine Intelligence  
We describe an object detection system based on mixtures of multiscale deformable part models.  ...  A latent SVM is semi-convex and the training problem becomes convex once latent information is specified for the positive examples.  ...  A latent SVM can be viewed as a type of energy-based model [27] .  ... 
doi:10.1109/tpami.2009.167 pmid:20634557 fatcat:bk4wloylvbfjfea4pzpbgnsrgm

Discriminative Neural Sentence Modeling by Tree-Based Convolution [article]

Lili Mou, Hao Peng, Ge Li, Yan Xu, Lu Zhang, Zhi Jin
2015 arXiv   pre-print
This paper proposes a tree-based convolutional neural network (TBCNN) for discriminative sentence modeling. Our models leverage either constituency trees or dependency trees of sentences.  ...  We evaluate our models on two tasks: sentiment analysis and question classification.  ...  Conclusion In this paper, we proposed a novel neural discriminative sentence model based on sentence parsing structures.  ... 
arXiv:1504.01106v5 fatcat:am4ydiibzra4zdnhyfnlovl5ly

Earth Mover's Distance-Based Local Discriminant Basis [chapter]

Bradley Marchand, Naoki Saito
2012 Multiscale Signal Analysis and Modeling  
Local discriminant Basis (LDB) is a tool to extract useful features for signal and image classification problems.  ...  Local Discriminant Basis (LDB) is a best basis algorithm developed by Saito and Coifman for the purpose of classification [9, 10] .  ...  To evaluate our algorithm on increasing classifier performance, we use two different base classifiers, Linear Discriminant Analysis (LDA) and Classification Tree (CT); see e.g., [4, Sec. 4 .3, Sec. 9.2  ... 
doi:10.1007/978-1-4614-4145-8_12 fatcat:zufzn42kwfcz5l5fwb3xd77mte

Variable selection in model-based clustering and discriminant analysis with a regularization approach [article]

Gilles Celeux, Cathy Maugis-Rabusseau, Mohammed Sedki
2017 arXiv   pre-print
In this paper, an alternative regularization approach of variable selection is proposed for model-based clustering and classification.  ...  Relevant methods of variable selection have been proposed in model-based clustering and classification.  ...  After a series of papers on variable selection in model-based clustering (Law et al, 2004; Tadesse et al, 2005; Raftery and Dean, 2006; Maugis et al, 2009a) , Maugis et al (2009b) proposed a general  ... 
arXiv:1705.00946v1 fatcat:k7mv4p6zmzderk27cp4atduupq

Model-based cluster and discriminant analysis with the MIXMOD software

Christophe Biernacki, Gilles Celeux, Gérard Govaert, Florent Langrognet
2006 Computational Statistics & Data Analysis  
The mixmod (mixture modeling) program fits mixture models to a given data set for the purposes of density estimation, clustering or discriminant analysis.  ...  analysis or discriminant analysis).  ...  The result should be compared with the solution obtained with the EM algorithm shown in Figure 2 (a). 5 Model-based discriminant analysis Data processed by mixmod for discriminant analysis consists of  ... 
doi:10.1016/j.csda.2005.12.015 fatcat:kcggwsw27bgorev2rx2pu42vue

Texture Discrimination Based Upon an Assumed Stochastic Texture Model

James W. Modestino, Robert W. Fries, Acie L. Vickers
1981 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Based upon this stochastic model we propose a new approach to texture discrimination which is an approximation to the statistically optimum maximum likelihood classifier.  ...  In the present paper we describe a class of 2-D random fields, of which the samples in Fig.'  ... 
doi:10.1109/tpami.1981.4767148 fatcat:66heupwwuzeczn7yd3g2qucmyi

Study on non-linear bistable dynamics model based EEG signal discrimination analysis method

Xiaoguo Ying, Han Lin, Guohua Hui
2015 Bioengineered  
In this paper, EEG signal discrimination based on non-linear bistable dynamical model was proposed.  ...  EEG signals were processed by non-linear bistable dynamical model, and features of EEG signals were characterized by coherence index.  ...  In this paper, EEG signal analysis based on non-linear bistable dynamical model was proposed.  ... 
doi:10.1080/21655979.2015.1065360 pmid:26176364 pmcid:PMC4825826 fatcat:gmho5jeuqbh63nzdw3tnem25by

Variable selection and updating in model-based discriminant analysis for high dimensional data with food authenticity applications

Thomas Brendan Murphy, Nema Dean, Adrian E. Raftery
2010 Annals of Applied Statistics  
Motivated by food authenticity applications, a model-based discriminant analysis method that includes variable selection is presented.  ...  The discriminant analysis model is fitted in a semi-supervised manner using both labeled and unlabeled data.  ...  We would like to thank the Editor, Associate Editor and Referees whose suggestions greatly improved this paper.  ... 
doi:10.1214/09-aoas279 pmid:20936055 pmcid:PMC2951685 fatcat:jnh5omxo65apxjiewqz2iojcnm

Variable selection in model-based clustering and discriminant analysis with a regularization approach

Gilles Celeux, Cathy Maugis-Rabusseau, Mohammed Sedki
2018 Advances in Data Analysis and Classification  
In this paper, we propose an alternative regularization approach for variable selection in model-based clustering and classification.  ...  In our approach the variables are first ranked using a lassolike procedure in order to avoid slow stepwise algorithms.  ...  After a series of papers on variable selection in model-based clustering (Law et al, 2004; Tadesse et al, 2005; Raftery and Dean, 2006; Maugis et al, 2009a) , Maugis et al (2009b) proposed a general  ... 
doi:10.1007/s11634-018-0322-5 fatcat:wjttmzvjkvb23ecqx6s2a6sake

Discriminator Augmented Model-Based Reinforcement Learning [article]

Behzad Haghgoo, Allan Zhou, Archit Sharma, Chelsea Finn
2021 arXiv   pre-print
By planning through a learned dynamics model, model-based reinforcement learning (MBRL) offers the prospect of good performance with little environment interaction.  ...  This paper aims to improve planning with an importance sampling framework that accounts and corrects for discrepancy between the true and learned dynamics.  ...  Fig. 2 depicts our analysis of what the trained model and discriminator are doing in IcyRoad: we see that the unimodal model is struggling to fit the true dynamics, while the discriminator is helping  ... 
arXiv:2103.12999v2 fatcat:7qrl3e4y7jb5jenmoeukudlzma

Responder Analyses—A PhRMA Position Paper

Tom Uryniak, Ivan S.F. Chan, Valerii V. Fedorov, Qi Jiang, Leonard Oppenheimer, Steven M. Snapinn, Chi-Hse Teng, John Zhang
2011 Statistics in Biopharmaceutical Research  
3, No. 3 DOI: 10.1198 /sbr.2011 Responder Analyses-A PhRMA Position Paper and regulatory guidance documents.  ...  Responder Analyses-A PhRMA Position Paper Responder Analyses and the Assessment of a Clinically Relevant Treatment Effect Perhaps the main problem with the responder analysis is the often arbitrary nature  ... 
doi:10.1198/sbr.2011.10070 fatcat:lgsqoqztzfdldasusqy6y5kbai
« Previous Showing results 1 — 15 out of 704,126 results