Shen-Shyang Ho

Micro-Doppler effect in radar: phenomenon, model, and simulation study

When, in addition to the constant Doppler frequency shift induced by the bulk motion of a radar target, the target or any structure on the target undergoes micro-motion dynamics, such as mechanical vibrations or rotations, the micro-motion dynamics induce Doppler modulations on the returned signal, referred to as the micro-Doppler effect. We introduce the micro-Doppler phenomenon in radar, develop a model of Doppler modulations, derive formulas of micro-Doppler induced by targets with vibration, rotation, tumbling and coning motions, and verify them by simulation studies, analyze time-varying micro-Doppler features using high-resolution time-frequency transforms, and demonstrate the micro-Doppler effect observed in real radar data.

A Martingale Framework for Detecting Changes in Data Streams by Testing Exchangeability

In a data streaming setting, data points are observed sequentially. The data generating model may change as the data are streaming. In this paper, we propose detecting this change in data streams by testing the exchangeability property of the observed data. Our martingale approach is an efficient, nonparametric, one-pass algorithm that is effective on the classification, cluster, and regression data generating models. Experimental results show the feasibility and effectiveness of the martingale methodology in detecting changes in the data generating model for time-varying data streams. Moreover, we also show that: (1) An adaptive support vector machine (SVM) utilizing the martingale methodology compares favorably against an adaptive SVM utilizing a sliding window, and (2) a multiple martingale video-shot change detector compares favorably against standard shot-change detection algorithms.

Differential privacy for location pattern mining

One main concern for individuals to participate in the data collection of personal location history records is the disclosure of their location and related information when a user queries for statistical or pattern mining results derived from these records. In this paper, we investigate how the privacy goal that the inclusion of ones location history in a statistical database with location pattern mining capabilities does not substantially increase ones privacy risk. In particular, we propose a differentially private pattern mining algorithm for interesting geographic location discovery using a region quadtree spatial decomposition to preprocess the location points followed by applying a density-based clustering algorithm. A differentially private region quadtree is used for both de-noising the spatial domain and identifying the likely geographic regions containing the interesting locations. Then, a differential privacy mechanism is applied to the algorithm outputs, namely: the interesting regions and their corresponding stay point counts. The quadtree spatial decomposition enables one to obtain a localized reduced sensitivity to achieve the differential privacy goal and accurate outputs. Experimental results on synthetic datasets are used to show the feasibility of the proposed privacy preserving location pattern mining algorithm.

An efficient regularized K-nearest neighbor based weighted twin support vector regression

Our RKNNWTSVR implements structural risk minimization principle by introducing extra regularization terms in each objective function.Our RKNNWTSVR cannot only help to alleviate overfitting issue and improve the generalization performance but also introduce invertibility in the dual formulation.The square of the 2-norm of the vector of slack variables is used in RKNNWTSVR to make the objective functions strongly convex.Four algorithms are designed to solve the proposed RKNNWTSVR.The solution reduces to solving just two systems of linear equations which makes our RKNNWTSVR extremely simple and efficient.No external optimizer is necessary for solving the RKNNWTSVR formulation. In general, pattern classification and regression tasks do not take into consideration the variation in the importance of the training samples. For twin support vector regression (TSVR), this implies that all the training samples play the same role on the bound functions. However, the number of close neighboring samples near to each training sample has an effect on the bound functions. In this paper, we formulate a regularized version of the KNN-based weighted twin support vector regression (KNNWTSVR) called RKNNWTSVR which is both efficient and effective. By introducing the regularization term and replacing 2-norm of slack variables instead of 1-norm, our RKNNWTSVR only needs to solve a simple system of linear equations with low computational cost, and at the same time, it improves the generalization performance. Particularly, we compare four implementations of RKNNWTSVR with existing approaches. Experimental results on several synthetic and benchmark datasets indicate that, comparing to SVR, WSVR, TSVR and KNNWTSVR, our proposed RKNNWTSVR has better generalization ability and requires less computational time.

Robust energy-based least squares twin support vector machines

Twin support vector machine (TSVM), least squares TSVM (LSTSVM) and energy-based LSTSVM (ELS-TSVM) satisfy only empirical risk minimization principle. Moreover, the matrices in their formulations are always positive semi-definite. To overcome these problems, we propose in this paper a robust energy-based least squares twin support vector machine algorithm, called RELS-TSVM for short. Unlike TSVM, LSTSVM and ELS-TSVM, our RELS-TSVM maximizes the margin with a positive definite matrix formulation and implements the structural risk minimization principle which embodies the marrow of statistical learning theory. Furthermore, RELS-TSVM utilizes energy parameters to reduce the effect of noise and outliers. Experimental results on several synthetic and real-world benchmark datasets show that RELS-TSVM not only yields better classification performance but also has a lower training time compared to ELS-TSVM, LSPTSVM, LSTSVM, TBSVM and TSVM.

ML-TREE: A Tree-Structure-Based Approach to Multilabel Learning

Multilabel learning aims to predict labels of unseen instances by learning from training samples that are associated with a set of known labels. In this paper, we propose to use a hierarchical tree model for multilabel learning, and to develop the ML-Tree algorithm for finding the tree structure. ML-Tree considers a tree as a hierarchy of data and constructs the tree using the induction of one-against-all SVM classifiers at each node to recursively partition the data into child nodes. For each node, we define a predictive label vector to represent the predictive label transmission in the tree model for multilabel prediction and automatic discovery of the label relationships. If two labels co-occur frequently as predictive labels at leaf nodes, these labels are supposed to be relevant. The amount of predictive label co-occurrence provides an estimation of the label relationships. We examine the ML-Tree method on 11 real data sets of different domains and compare it with six well-established multilabel learning algorithms. The performances of these approaches are evaluated by 16 commonly used measures. We also conduct Friedman and Nemenyi tests to assess the statistical significance of the differences in performance. Experimental results demonstrate the effectiveness of our method.

Collective prediction of protein functions from protein-protein interaction networks

BackgroundAutomated assignment of functions to unknown proteins is one of the most important task in computational biology. The development of experimental methods for genome scale analysis of molecular interaction networks offers new ways to infer protein function from protein-protein interaction (PPI) network data. Existing techniques for collective classification (CC) usually increase accuracy for network data, wherein instances are interlinked with each other, using a large amount of labeled data for training. However, the labeled data are time-consuming and expensive to obtain. On the other hand, one can easily obtain large amount of unlabeled data. Thus, more sophisticated methods are needed to exploit the unlabeled data to increase prediction accuracy for protein function prediction.ResultsIn this paper, we propose an effective Markov chain based CC algorithm (ICAM) to tackle the label deficiency problem in CC for interrelated proteins from PPI networks. Our idea is to model the problem using two distinct Markov chain classifiers to make separate predictions with regard to attribute features from protein data and relational features from relational information. The ICAM learning algorithm combines the results of the two classifiers to compute the ranks of labels to indicate the importance of a set of labels to an instance, and uses an ICA framework to iteratively refine the learning models for improving performance of protein function prediction from PPI networks in the paucity of labeled data.ConclusionExperimental results on the real-world Yeast protein-protein interaction datasets show that our proposed ICAM method is better than the other ICA-type methods given limited labeled training data. This approach can serve as a valuable tool for the study of protein function prediction from PPI networks.

Semi-supervised multi-label collective classification ensemble for functional genomics

BackgroundWith the rapid accumulation of proteomic and genomic datasets in terms of genome-scale features and interaction networks through high-throughput experimental techniques, the process of manual predicting functional properties of the proteins has become increasingly cumbersome, and computational methods to automate this annotation task are urgently needed. Most of the approaches in predicting functional properties of proteins require to either identify a reliable set of labeled proteins with similar attribute features to unannotated proteins, or to learn from a fully-labeled protein interaction network with a large amount of labeled data. However, acquiring such labels can be very difficult in practice, especially for multi-label protein function prediction problems. Learning with only a few labeled data can lead to poor performance as limited supervision knowledge can be obtained from similar proteins or from connections between them. To effectively annotate proteins even in the paucity of labeled data, it is important to take advantage of all data sources that are available in this problem setting, including interaction networks, attribute feature information, correlations of functional labels, and unlabeled data.ResultsIn this paper, we show that the underlying nature of predicting functional properties of proteins using various data sources of relational data is a typical collective classification (CC) problem in machine learning. The protein functional prediction task with limited annotation is then cast into a semi-supervised multi-label collective classification (SMCC) framework. As such, we propose a novel generative model based SMCC algorithm, called GM-SMCC, to effectively compute the label probability distributions of unannotated protein instances and predict their functional properties. To further boost the predicting performance, we extend the method in an ensemble manner, called EGM-SMCC, by utilizing multiple heterogeneous networks with various latent linkages constructed to explicitly model the relationships among the nodes for effectively propagate the supervision knowledge from labeled to unlabeled nodes.ConclusionExperimental results on a yeast gene dataset predicting the functions and localization of proteins demonstrate the effectiveness of the proposed method. In the comparison, we find that the performances of the proposed algorithms are better than the other compared algorithms.

Tropical cyclone event sequence similarity search via dimensionality reduction and metric learning

The Earth Observing System Data and Information System (EOSDIS) is a comprehensive data and information system which archives, manages, and distributes Earth science data from the EOS spacecrafts. One non-existent capability in the EOSDIS is the retrieval of satellite sensor data based on weather events (such as tropical cyclones) similarity query output. In this paper, we propose a framework to solve the similarity search problem given user-defined instance-level constraints for tropical cyclone events, represented by arbitrary length multidimensional spatio-temporal data sequences. A critical component for such a problem is the similarity/metric function to compare the data sequences. We describe a novel Longest Common Subsequence (LCSS) parameter learning approach driven by nonlinear dimensionality reduction and distance metric learning. Intuitively, arbitrary length multidimensional data sequences are projected into a fixed dimensional manifold for LCSS parameter learning. Similarity search is achieved through consensus among the (similar) instance-level constraints based on ranking orders computed using the LCSS-based similarity measure. Experimental results using a combination of synthetic and real tropical cyclone event data sequences are presented to demonstrate the feasibility of our parameter learning approach and its robustness to variability in the instance constraints. We, then, use a similarity query example on real tropical cyclone event data sequences from 2000 to 2008 to discuss (i) a problem of scientific interest, and (ii) challenges and issues related to the weather event similarity search problem.

Automated Cyclone Identification From Remote QuikSCAT Satellite Data

We discuss a fully automated remote cyclone identification and tracking approach using the QuikSCAT wind sensor data. Our approach consists of five main automated steps: QuikSCAT data retrieval, QuikSCAT feature extraction & data preprocessing, cyclone identification, motion/location prediction, and cyclone tracking. Ensemble learning based on a committee of support vector machines using features extracted from QuikSCAT wind sensor data are used for cyclone identification. Experimental results demonstrates the feasibility and usefulness of our automated approach.