Incremental Support Vector Regression Research Articles

Classification and analysis of cardiac arrhythmia based on incremental support vector regression on IOT platform

The electrocardiogram (ECG) is a diagnostic device capable of monitoring normal or irregular heart function. The entire ECG beat can be categorized into five different forms of beat arrhythmias (N, S, V, F, U). Quick and precise diagnosis of forms of arrhythmia is critical for identifying the heart problem and provides the proper treatment to the patient. In this paper, Discrete Wavelet Transform and Higher Order Statistics techniques has been used for analyzing and determining the ECG signals and implement it on an IOT-based platform. This system is based on three categories: The first approach involves inputting the ECG data; the second approach involves extracting the ECG beats with their respective amplitude from the base line. Wavelet transform function, and higher order statistics are used to eliminate noise and unwanted signal components and thus to extract ECG features. The third approach is to classify the ECG beats based on the Incremental Support vector regression classifier. After classification ECG beat is transmitted to the controller section for signal processing are given to controller section (Arduino Uno). The process can be implemented by employing the statistical feature for the feature extraction from the ECG signal. Compared to other approaches, the method provided by Incremental Support vector regression to identify the ECG beats and predict arrhythmia can provide successful detection of arrhythmias. The basic concept of the proposed system is to provide patients with reliable health care by using cloud data compliance to allow doctors to use this information and to provide a fast and feasible service. The findings show that the proposed algorithm is successful in predicting cardiac arrhythmias, with a 98% that is higher than other approaches.

Application of incremental support vector regression based on optimal training subset and improved particle swarm optimization algorithm in real-time sensor fault diagnosis

Attracted by the advantages of support vector regression and incremental learning approach, it is proposed in this work that an incremental support vector regression (ISVR) model optimized by particle swarm optimization (PSO) algorithm, and some improvements are made to be more suitable for sensor faults on-line diagnosis. To reducethe training time of ISVR model, an optimal training subset (OTS) method is adopted to reduce the size of training data set of the model. Then, in order to solve the problem of slow convergence of standard PSO algorithm, an incremental PSO (IPSO) algorithm is proposed to accelerate the model convergence through adjusting the inertial weight of each particle, which is gained by comparing the current position of each particle and the optimal position of the last incremental training. Based on the above improvements, a hybrid model, IPSO-OTS-ISVR model is presented finally. Experimental results based on actual operational data of a gas turbine shows that, under the premise of ensuring accuracy, the proposed IPSO-OTS-ISVR has much better performance in model response time and convergence performance over the comparison models. The experimental results based on an UCI data set indicate that the proposed hybrid model can also be extended to solve other prediction problems.

Reliable state of health condition monitoring of Li-ion batteries based on incremental support vector regression with parameters optimization

State of health condition monitoring of Li-ion batteries is an important issue for safe and reliably operation of battery-powered products. Consequently, it remains a challenging subject for industrial and academic studies. In this article, an incremental support vector regression is proposed for battery state of health lifetime estimation. In order to improve the battery state of health forecasting accuracy, the quantum-behaved particle swarm optimization is proposed to define reliably the incremental support vector regression parameters. The validation of the proposed method was done based on the NASA battery data set, and it demonstrates that it yields good performance in remaining useful life estimation of Li-ion batteries. This case study shows that compared with the linear, polynomial regression methods, and compared to previous works, the proposed method can obtain more accurate state of health prediction results. Even for state of health prediction starting from the cycle near capacity regeneration, the proposed model can still accurately estimate the global degradation trend. Furthermore, the proposed quantum-behaved particle swarm optimization–incremental support vector regression combination has greater robustness when the training data contain noise and measurement outliers. This allows satisfactory prediction performances without pre-processing the data manually.

A new suitable feature selection and regression procedure for lithium-ion battery prognostics

The accurate prediction of lithium-ion battery Remaining Useful Life (RUL) is indispensable for safe and lifetime-optimised operation. Thereby, the monitoring of this vital component is very necessary for planning repair work and minimising unexpected electricity outage. However, the study and the investigation of internal battery parameters show several value changes within the battery lifetime and it is highly influenced by environmental and load conditions. Consequently, this paper presents a new prognostic method for online battery monitoring based on isometric feature mapping technique (ISOMAP) and incremental support vector regression (ISVR). ISOMAP is used to reduce some features extracted from lithium-ion batteries, with different health states, in both modes of charge and discharge, and ISVR is used to regress online the selected feature. Experimental results show that the proposed methodology provides a new suitable trend parameter for battery RUL prediction.

Bellman residuals minimization using online support vector machines

In this paper we present and theoretically study an Approximate Policy Iteration (API) method called A P I − B R M 𝜖 using a very effective implementation of incremental Support Vector Regression (SVR) to approximate the value function able to generalize Reinforcement Learning (RL) problems with continuous (or large) state space. A P I − B R M 𝜖 is presented as a non-parametric regularization method based on an outcome of the Bellman Residual Minimization (BRM) able to minimize the variance of the problem. The proposed method can be cast as incremental and may be applied to the on-line agent interaction framework of RL. Being also based on SVR which are based on convex optimization, is able to find the global solution of the problem. A P I − B R M 𝜖 using SVR can be seen as a regularization problem using 𝜖−insensitive loss. Compared to standard squared loss also used in regularization, this allows to naturally build a sparse solution for the approximation function. We extensively analyze the statistical properties of A P I − B R M 𝜖 founding a bound which controls the performance loss of the algorithm under some assumptions on the kernel and assuming that the collected samples are not-i.i.d. following a β−mixing process. Some experimental evidence and performance for well known RL benchmarks are also presented.

An incremental electric load forecasting model based on support vector regression

With the smart portable systems and the daily growth of databases on the web, there are ever-increasing requirements to learn the batch arriving and large sample data set. In this paper, an incremental learning model of support vector regression (SVR) is proposed to forecast the electric load under the batch arriving and large sample. For modeling with SVR, the optimal embedding of time series is constructed by phase space reconstruction (PSR). Then, an optimal training subset for the training of SVR is extracted based on the current data set, which enables us to cut the high time and space complexity by reducing the full training data set. When newly-increased data are added into the system, a representative data set reconstruction method is presented for quickly re-training the current SVR, and a nested particle swarm optimization (NPSO) framework is presented to select the parameters of the incremental SVR model. Experiments of incremental electric load forecasting demonstrate the computational superiority of the presented model over the comparison models.

Approximate policy iteration using regularised Bellman residuals minimisation

In this paper we present an approximate policy iteration (API) method called API‐BRMϵ using a very effective implementation of incremental support vector regression (SVR) to approximate the value function able to generalise in continuous (or large) space reinforcement learning (RL) problems. RL is a methodology able to solve complex and uncertain decision problems usually modelled as Markov decision problems. API-BRMϵ is formalised as a non-parametric regularisation problem based on an outcome of the Bellman residual minimisation (BRM) which is able to minimise the variance of the problem. API-BRMϵ is incremental and can be applied to RL using the on-line agent interaction framework. Based on non-parametric SVR API-BRMϵ is able to find the global solution of the problem with convergence guarantees to the optimal solution. A value function should be defined to find the optimal policy specifying the total reward that an agent might expect in its current state taking one action. Therefore, the agent will use the value function to choose the action to take. Some experimental evidence and performance for well-known RL benchmarks are presented.

Efficient Data Capturing for Network Forensics in Cognitive Radio Networks

Network forensics is an emerging interdiscipline used to track down cyber crimes and detect network anomalies for a multitude of applications. Efficient capture of data is the basis of network forensics. Compared to traditional networks, data capture faces significant challenges in cognitive radio networks. In traditional wireless networks, usually one monitor is assigned to one channel for traffic capture. This approach will incur very high cost in cognitive radio networks because it typically has a large number of channels. Furthermore, due to the uncertainty of the primary user's behavior, cognitive radio devices change their operating channels dynamically, which makes data capturing more difficult. In this paper, we propose a systematic method to capture data in cognitive radio networks with a small number of monitors. We utilize incremental support vector regression to predict packet arrival time and intelligently switch monitors between channels. We also propose a protocol that schedules multiple monitors to perform channel scanning and packet capturing in an efficient manner. Monitors are reused in the time domain, and geographic coverage is taken into account. The real-world experiments and simulations show that our method is able to achieve the packet capture rate above 70% using a small number of monitors, which outperforms the random scheme by 200%-300%.

Incremental smooth support vector regression for Takagi–Sugeno fuzzy modeling

We propose an architecture for Takagi–Sugeno (TS) fuzzy system and develop an incremental smooth support vector regression (ISSVR) algorithm to build the TS fuzzy system. ISSVR is based on the ε-insensitive smooth support vector regression (ε-SSVR), a smoothing strategy for solving ε-SVR, and incremental reduced support vector machine (RSVM). The ISSVR incrementally selects representative samples from the given dataset as support vectors. We show that TS fuzzy modeling is equivalent to the ISSVR problem under certain assumptions. A TS fuzzy system can be generated from the given training data based on the ISSVR learning with each fuzzy rule given by a support vector. Compared with other fuzzy modeling methods, more forms of membership functions can be used in our model, and the number of fuzzy rules of our model is much smaller. The performance of our model is illustrated by extensive experiments and comparisons.

FAST FUZZY MODELING USING A NEW INCREMENTAL SVR

In this paper, we propose a novel approach to fast fuzzy modeling based on a new incremental support vector regression (SVR). Firstly a candidate support vectors selection strategy based on kernel Mahalanobis distance measurement is proposed. This strategy is further used to develop a new incremental learning algorithm to speed up the training process of SVR. Then a hybrid kernel function is utilized to represent an SVR model as a TS fuzzy model. Finally a set of fuzzy rules can be directly extracted from the learning results of SVR. Experimental results of two benchmark examples show that the proposed model not only possesses satisfactory accuracy and generalization ability but also costs less computational time.

Support vector regression for on-line health monitoring of large-scale structures

Large-scale, structural health monitoring remains a challenge especially when I/O measurement data are contaminated by high-level noise. A novel approach that uses incremental support vector regression (SVR), a promising statistics technology, is proposed for large-scale, structural health monitoring. Due to the potential properties of this novel SVR, the SVR-based approach makes structural health monitoring accurately and robustly. A sub-structure strategy is utilized to reduce the number of unknown parameters in the health monitoring formula, thereby making large-scale structural health monitoring possible. Lastly, an incremental SVR training algorithm adopted for the SVR-based approach not only markedly reduces computation time, but identifies structural parameters on-line. Numerical examples show that results of this SVR-based approach for large-scale structural health monitoring are accurate and robust, even when observed data are contaminated with different kinds and intensity levels of noise.

NON-LINEAR HYSTERETIC STRUCTURAL IDENTIFICATION BY UTILIZING ON-LINE SUPPORT VECTOR REGRESSION

Structures exhibit highly nonlinear characters under severe loads such as strong seismic excitations. Therefore, it is crucial to make nonlinear structural identification in civil engineering. However, nonlinear hysteretic structural identification is still a challenging topic due to structural model complexity and the strong noises existing in input and output (I/O) data. An efficient approach based on the incremental support vector regression (SVR) is proposed here to identify nonlinear hysteretic structural parameters on-line. Instead of the Gaussian loss function utilized in the least squares method, a novel einsensitive loss function is employed in SVR, and therefore the suggested SVR-based approach produces robust and accurate identification results. Furthermore, as an incremental algorithm employed to train SVR in a sequential way, the presented SVR-based approach not only works rapidly, but also identifys nonlinear structural constitutive parameters on-line. The performance of the proposed approach is verified by a five degree of freedom nonlinear hysteretic structural identification problem, in which two cases (power parameter is known/unknown) are both investigated. The identified results show evidently that the proposed technique has potential performance in robustness and accuracy for nonlinear structural identification, even when the measurement data in the presence of noises.

An Incremental Learning Strategy for Support Vector Regression

Support vector machine (SVM) provides good generalization performance but suffers from a large amount of computation. This paper presents an incremental learning strategy for support vector regression (SVR). The new method firstly formulates an explicit expression of ||W||2 by constructing an orthogonal basis in feature space together with a basic Hilbert space identity, and then finds the regression function through minimizing the formula of ||W||2 rather than solving a convex programming problem. Particularly, we combine the minimization of ||W||2 with kernel selection that can lead to good generalization performance. The presented method not only provides a novel way for incremental SVR learning, but opens an opportunity for model selection of SVR as well. An artificial data set, a benchmark data set and a real-world data set are employed to evaluate the method. The simulations support the feasibility and effectiveness of the proposed approach.