Hong Man

DCPE co-training for classification

Co-training is a well-known semi-supervised learning technique that applies two basic learners to train the data source, which uses the most confident unlabeled data to augment labeled data in the learning process. In the paper, we use the diversity of class probability estimation (DCPE) between two learners and propose the DCPE co-training approach. The key idea is to use DCPE to predict labels for the unlabeled data in the training process. The experimental studies with UCI data demonstrate that the DCPE co-training is robust and efficient in classification. The comparative studies with supervised learning methods and semi-supervised learning methods also demonstrate the effectiveness of the proposed approach.

SOMKE: Kernel Density Estimation Over Data Streams by Sequences of Self-Organizing Maps

In this paper, we propose a novel method SOMKE, for kernel density estimation (KDE) over data streams based on sequences of self-organizing map (SOM). In many stream data mining applications, the traditional KDE methods are infeasible because of the high computational cost, processing time, and memory requirement. To reduce the time and space complexity, we propose a SOM structure in this paper to obtain well-defined data clusters to estimate the underlying probability distributions of incoming data streams. The main idea of this paper is to build a series of SOMs over the data streams via two operations, that is, creating and merging the SOM sequences. The creation phase produces the SOM sequence entries for windows of the data, which obtains clustering information of the incoming data streams. The size of the SOM sequences can be further reduced by combining the consecutive entries in the sequence based on the measure of Kullback-Leibler divergence. Finally, the probability density functions over arbitrary time periods along the data streams can be estimated using such SOM sequences. We compare SOMKE with two other KDE methods for data streams, the M-kernel approach and the cluster kernel approach, in terms of accuracy and processing time for various stationary data streams. Furthermore, we also investigate the use of SOMKE over nonstationary (evolving) data streams, including a synthetic nonstationary data stream, a real-world financial data stream and a group of network traffic data streams. The simulation results illustrate the effectiveness and efficiency of the proposed approach.

Sparse-Representation-Based Classification with Structure-Preserving Dimension Reduction

Sparse-representation-based classification (SRC), which classifies data based on the sparse reconstruction error, has been a new technique in pattern recognition. However, the computation cost for sparse coding is heavy in real applications. In this paper, various dimension reduction methods are studied in the context of SRC to improve classification accuracy as well as reduce computational cost. A feature extraction method, i.e., principal component analysis, and feature selection methods, i.e., Laplacian score and Pearson correlation coefficient, are applied to the data preparation step to preserve the structure of data in the lower-dimensional space. Classification performance of SRC with structure-preserving dimension reduction (SRC–SPDR) is compared to classical classifiers such as k-nearest neighbors and support vector machines. Experimental tests with the UCI and face data sets demonstrate that SRC–SPDR is effective with relatively low computation cost

Feature selection based on sparse imputation

Feature selection, which aims to obtain valuable feature subsets, has been an active topic for years. How to design an evaluating metric is the key for feature selection. In this paper, we address this problem using imputation quality to search for the meaningful features and propose feature selection via sparse imputation (FSSI) method. The key idea is utilizing sparse representation criterion to test individual feature. The feature based classification is used to evaluate the proposed method. Comparative studies are conducted with classic feature selection methods (such as Fisher score and Laplacian score). Experimental results on benchmark data sets demonstrate the effectiveness of FSSI method.

Spatial outlier detection based on iterative self-organizing learning model

In this paper, we propose an iterative self-organizing map (SOM) approach with robust distance estimation (ISOMRD) for spatial outlier detection. Generally speaking, spatial outliers are irregular data instances which have significantly distinct non-spatial attribute values compared to their spatial neighbors. In our proposed approach, we adopt SOM to preserve the intrinsic topological and metric relationships of the data distribution to seek reasonable spatial clusters for outlier detection. The proposed iterative learning process with robust distance estimation can address the high dimensional problems of spatial attributes and accurately detect spatial outliers with irregular features. To verify the efficiency and robustness of our proposed algorithm, comparative study of ISOMRD and several existing approaches are presented in detail. Specifically, we test the performance of our method based on four real-world spatial datasets. Various simulation results demonstrate the effectiveness of the proposed approach.

SOMSO: A self-organizing map approach for spatial outlier detection with multiple attributes

In this paper, we propose a self-organizing map approach for spatial outlier detection, the SOMSO method. Spatial outliers are abnormal data points which have significantly distinct non-spatial attribute values compared with their neighborhood. Detection of spatial outliers can further discover spatial distribution and attribute information for data mining problems. Self-Organizing map (SOM) is an effective method for visualization and cluster of high dimensional data. It can preserve intrinsic topological and metric relationships in datasets. The SOMSO method can solve high dimensional problems for spatial attributes and accurately detect spatial outliers with irregular features. The experimental results for the dataset based on U.S. population census indicate that SOMSO approach can successfully be applied in complicated spatial datasets with multiple attributes.

A General-Purpose FPGA-Based Reconfigurable Platform for Video and Image Processing

This paper presents a general-purpose, multi-task, and reconfigurable platform for video and image processing. With the increasing requirements of processing power in many of todays video and image processing applications, it is important to go beyond the software implementation to provide a real-time, low cost, high performance, and scalable hardware platform. In this paper, we propose a system by using the powerful parallel processing architecture in the Field Programmable Gate Array (FPGA) to achieve this objective. Based on the proposed system level architecture and design strategies, a prototype system is developed based on the Xilinx Virtex-II FPGA with the integration of embedded processor, memory control and interface technologies. Our system includes different functional modules, such as edge detection, zoom-in and zoom-out functions, which provides the flexibility of using this system as a general video processing platform according to different application requirements. The final system utilizes about 20% of logic resource, 50% of memory on chip, and has total power consumption around 203 mw.

IterativeSOMSO: an iterative self-organizing map for spatial outlier detection

In this paper, we propose an iterative self-organizing map approach for spatial outlier detection (IterativeSOMSO) IterativeSOMSO method can address high dimensional problems for spatial attributes and accurately detect spatial outliers with irregular features Detection of spatial outliers facilitates further discovery of spatial distribution and attribute information for data mining problems The experimental results indicate our proposed approach can be effectively implemented for the large spatial dataset based on U.S Census Bureau with approving performance.

DCPE co-training: Co-training based on diversity of class probability estimation

Co-training is a semi-supervised learning technique used to recover the unlabeled data based on two base learners. The normal co-training approaches use the most confidently recovered unlabeled data to augment the training data. In this paper, we investigate the co-training approaches with a focus on the diversity issue and propose the diversity of class probability estimation (DCPE) co-training approach. The key idea of the DCPE co-training method is to use DCPE between two base learners to choose the recovered unlabeled data. The results are compared with classic co-training, tri-training and self training methods. Our experimental study based on the UCI benchmark data sets shows that the DCPE co-training is robust and efficient in the classification.

Hybrid learning based on Multiple Self-Organizing Maps and Genetic Algorithm

Multiple Self-Organizing Maps (MSOMs) based classification methods are able to combine the advantages of both unsupervised and supervised learning mechanisms. Specifically, unsupervised SOM can search for similar properties from input data space and generate data clusters within each class, while supervised SOM can be trained from the data via label matching in the global SOM lattice space. In this work, we propose a novel classification method that integrates MSOMs with Genetic Algorithm (GA) to avoid the influence of local minima. Davies-Bouldin Index (DBI) and Mean Square Error (MSE) are adopted as the objective functions for searching the optimal solution space. Experimental results demonstrate the effectiveness and robustness of our proposed approach based on several benchmark data sets from UCI Machine Learning Repository.