Hongming Zhou

Extreme Learning Machine for Regression and Multiclass Classification

Due to the simplicity of their implementations, least square support vector machine (LS-SVM) and proximal support vector machine (PSVM) have been widely used in binary classification applications. The conventional LS-SVM and PSVM cannot be used in regression and multiclass classification applications directly, although variants of LS-SVM and PSVM have been proposed to handle such cases. This paper shows that both LS-SVM and PSVM can be simplified further and a unified learning framework of LS-SVM, PSVM, and other regularization algorithms referred to extreme learning machine (ELM) can be built. ELM works for the “generalized” single-hidden-layer feedforward networks (SLFNs), but the hidden layer (or called feature mapping) in ELM need not be tuned. Such SLFNs include but are not limited to SVM, polynomial network, and the conventional feedforward neural networks. This paper shows the following: 1) ELM provides a unified learning platform with a widespread type of feature mappings and can be applied in regression and multiclass classification applications directly; 2) from the optimization method point of view, ELM has milder optimization constraints compared to LS-SVM and PSVM; 3) in theory, compared to ELM, LS-SVM and PSVM achieve suboptimal solutions and require higher computational complexity; and 4) in theory, ELM can approximate any target continuous function and classify any disjoint regions. As verified by the simulation results, ELM tends to have better scalability and achieve similar (for regression and binary class cases) or much better (for multiclass cases) generalization performance at much faster learning speed (up to thousands times) than traditional SVM and LS-SVM.

Robust Extreme Learning Machine With its Application to Indoor Positioning

The increasing demands of location-based services have spurred the rapid development of indoor positioning system and indoor localization system interchangeably (IPSs). However, the performance of IPSs suffers from noisy measurements. In this paper, two kinds of robust extreme learning machines (RELMs), corresponding to the close-to-mean constraint, and the small-residual constraint, have been proposed to address the issue of noisy measurements in IPSs. Based on whether the feature mapping in extreme learning machine is explicit, we respectively provide random-hidden-nodes and kernelized formulations of RELMs by second order cone programming. Furthermore, the computation of the covariance in feature space is discussed. Simulations and real-world indoor localization experiments are extensively carried out and the results demonstrate that the proposed algorithms can not only improve the accuracy and repeatability, but also reduce the deviation and worst case error of IPSs compared with other baseline algorithms.

Credit risk evaluation with extreme learning machine

Credit risk evaluation has become an increasingly important field in financial risk management for financial institutions, especially for banks and credit card companies. Many data mining and statistical methods have been applied to this field. Extreme learning machine (ELM) classifier as a type of generalized single hidden layer feed-forward networks has been used in many applications and achieve good classification accuracy. Thus, we use ELM (kernel based) as a classification tool to perform the credit risk evaluation in this paper. The simulations are done on two credit risk evaluation datasets with three different kernel functions. Simulation results show that the kernel based ELM is more suitable for credit risk evaluation than the popular used Support Vector Machines (SVMs) with consideration of overall, good and bad accuracies.

Physical field estimation from CFD database and sparse sensor observations

This paper presents a new approach to estimate one physical field from off-line computational fluid dynamics (CFD) database and real-time sparse sensor observations. Firstly, we determine the proper orthogonal decomposition (POD) modes from the CFD database. Then, we use extreme learning machine (ELM) to build a regression model between the boundary conditions of physical fields and their POD coefficients. With this model, we can directly estimate the physical field of interest. Next, we modify the estimated physical field based on sparse sensor observations with the help of the dominant POD modes. The modified physical field is shown more accurate than the physical field estimated from either the regression model or sensor observations. Finally, we provide a simple example to show the effectiveness of the proposed approach.

ELM Based Fast CFD Model with Sensor Adjustment

Computational fluid dynamics (CFD) simulation is a useful tool to provide temperature and air flow patterns of an indoor environment. However, one big drawback of the CFD simulation is that it usually requires high computational power and takes long time for one simulation. Thus it is hard to apply the CFD results in the HVAC real-time dynamic visualization system. In this paper, we proposed a fast machine learning approach to solve the CFD problem. The extreme learning machine (ELM) is used to learn the CFD model and it can estimate the CFD results approximately 100 times faster than the conventional CFD simulator. The estimated results could be further calibrated based on the indoor sensor information. ELM is also helpful to generate the error surface to correct the previous estimated CFD surface. The whole system only takes a few seconds to complete the CFD task in our case and thus it is possible to achieve the real-time visualization purpose.

Compressed representation learning for fluid field reconstruction from sparse sensor observations

This paper provides a new approach to reconstruct a fluid field from sparse sensor observations. Using the extreme learning machine (ELM) autoencoder, we can extract a dominant basis of the fluid field of interest from a database consisting of a series of fluid field snapshots obtained from offline computational fluid dynamics (CFD) simulations. The output weights of ELM autoencoder can be viewed as the compressed feature representations of the fluid field and represent the dominant behaviors of the database. With such a compressed representation, the fluid field of interest can be easily reconstructed from sparse sensor observations. The simulation results show that the new compressed representation approach can achieve better reconstruction accuracy as compared with the traditional principal component analysis (PCA) method.