In a previous paper a new version of nonlinear dimensionality reduction algorithm was proposed, the SC-LLE approach. This approach combines a supervised method, linear discriminant analysis (LDA, a simple but widely u...
详细信息
ISBN:
(纸本)9781467353076;9781467353069
In a previous paper a new version of nonlinear dimensionality reduction algorithm was proposed, the SC-LLE approach. This approach combines a supervised method, linear discriminant analysis (LDA, a simple but widely used algorithm in patternrecognition) with an unsupervised method, local linear embedding (LLE, manifold learning). SC-LLE method can generalize any linear classifier (like LDA) to nonlinear by transforming data into some low-dimensional feature space. This new concept (SC-LLE) applied to nonlinear data projection seems to be promising, and we show in this new paper that semi-supervised learning (SSL) is another interesting property of SC-LLE. Applications on 3D data show the interest of this method.
Anomalies in computer networks has increased in the last decades and raised concern to create techniques to identify these unusual traffic patterns. This research aims to use datamining techniques in order to correct...
详细信息
The poor performance and the lack of manual labeled corpus are two main problems in the task of protein-protein interaction extraction. A novel hybrid method is proposed. Based on the individual characteristics of mac...
详细信息
ISBN:
(纸本)9780769549231
The poor performance and the lack of manual labeled corpus are two main problems in the task of protein-protein interaction extraction. A novel hybrid method is proposed. Based on the individual characteristics of machinelearning and patternlearning, this method utilizes learned patterns from patternlearning to generate pattern features by performing sequence alignment. The pattern features and word features are incorporated into the input feature set of machinelearning algorithms. The semi-supervised method based on k-nearest neighbours classifier is also proposed to train the hybrid method from unlabeled data automatically. Experimental results show the improved performance over the baseline methods with the hybrid model and the efficieny of the semi-supervised method for the lack of labeled data.
The proceedings contain 49 papers. The topics discussed include: the gapped spectrum kernel for support vector machines;typhoon damage scale forecasting with self-organizing maps trained by selective presentation lear...
ISBN:
(纸本)9783642397110
The proceedings contain 49 papers. The topics discussed include: the gapped spectrum kernel for support vector machines;typhoon damage scale forecasting with self-organizing maps trained by selective presentation learning;dynamic-radius species-conserving genetic algorithm for the financial forecasting of Dow Jones index stocks;multi model transfer learning with RULES family;3D geovisualization techniques applied in spatial datamining;improving the efficiency of distributed datamining using an adjustment work flow;sign language recognition with support vector machines and hidden conditional random fields: going from fingerspelling to natural articulated words;classification and outlier detection based on topic based pattern synthesis;decremental learning of evolving fuzzy inference systems: application to handwritten gesture recognition;and unsupervised tagging of Spanish lyrics dataset using clustering.
The proceedings contain 55 papers. The topics discussed include: simitar: simplified searching of statistically significant similarity structure;second order scattering descriptors predict fMRI activity due to visual ...
ISBN:
(纸本)9780769550619
The proceedings contain 55 papers. The topics discussed include: simitar: simplified searching of statistically significant similarity structure;second order scattering descriptors predict fMRI activity due to visual textures;learning predictive cognitive structure from fMRI using supervised topic models;deriving statistical significance maps for support vector regression using medical imaging data;robust group-level inference in neuroimaging genetic studies;creating group-level functionally-defined atlases for diagnostic classification;clustering of high dimensional longitudinal imaging data;information criteria for dynamic contrast-enhanced magnetic resonance imaging;fast clustering for interactive tractography segmentation;importance sampling spherical harmonics to improve probabilistic tractography;comparing structural brain connectivity by the infinite relational model;and mining the hierarchy of resting-state brain networks: selection of representative clusters in a multiscale structure.
LLE is a very effective non-linear dimension reduction algorithm and widely explored in machinelearning, patternrecognition, datamining and etc. 'Locally linear, Globally non-linear' has always been regarde...
详细信息
Determining the accuracy and duration of AI techniques is important for the reason that the Application of such techniques in datamining has become wider. There is a range of techniques used in AI for finding hidden ...
详细信息
Rule-based systems like learning Classifier System are widely used in areas where datamining, data classification and patternrecognition tasks are essential. It is often difficult to address the knowledge base of th...
详细信息
Kernel methods play an important role in machinelearning, patternrecognition and datamining. Although the kernel functions are the central part of the kernel methods, little is known about the structure of its repr...
详细信息
Classification of network traffic for intrusion detection is a Big data classification problem. It requires an efficient machinelearning technique to learn the characteristics of the rapidly changing varieties of tra...
详细信息
暂无评论