The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Large and complex problem can be solved easily and quickly by decomposing it to be small sub-problems. We propose a heuristic method to isolate the larger state space into some smaller state spaces for decomposing learning task. During the learning process, after remove the state loops in these learned episodes, we find some states are critical for agent can reach goal state. These critical states...
Based on chaotic characteristic of high frequency ground-wave radar (HFGWR) sea clutter, a new adaptive artificial neural networks ensemble method for sea clutter predicting is presented in this paper. In phase space reconstructed, when one sea clutter sample is to be predicted, some artificial neural networks are choosed adaptively by evaluating their performance and error correlation in neighborhood...
Location and tracking the human faces is one of the critical technologies in free stereoscopic display system. But because of illumination variation or facial expression, it is difficult to detect human faces accurately and fast. In this paper, an infrared face detection based on real Adaboost algorithm and Cascade structure is implemented. With active infrared illumination, the problem caused by...
Transductive inference based on support vector machine is a new research region in statistical learning theory. An improved algorithm is proposed in this paper, which overcome the disadvantages of studying process complexity and slow in the progressive transductive support vector machine learning algorithm. The algorithm optimized the samples which near the support vector only, and large number of...
The prediction for pests is usually amphibious, relevant, complicated, and nonlinear. The neural network has the problem of decreasing generalization ability in the prediction of small samples. This paper presents a method of the prediction of pests based on fuzzy RBF neural network. A learning algorithm of adjusting the center, width and weight of the RBF is put forward. The use of fuzzy clustering...
To deal with unknown word and segmentation ambiguity, segmentation rules and tri-gram was used in inductive learning method. Rules were used for elementary segmentation and for better processing effectiveness in following steps. Those rules were acquired by manual labor through analyzing a tagged corpus. Inductive learning method recognized, extracted the unknown words from segmentation text recursively...
Learning machine is usually divided to strong learning machines and weak learning machines in machine learning. The result of most individual learning machine is output as while learning machine integration to used for a classification. BAN is an augmented Bayesian network classifier, whose accuracy can be improve by combining several weak learning machines. In this paper, a bagging classifier bagging-BAN-GBN...
A reliable and accurate short-term traffic forecasting system is crucial for the successful deployment of any intelligent transportation system. To address the complexity of real-world traffic forecasting conditions, this paper presents a layered traffic forecasting algorithm, which is implemented by a clustering neural network, Kohonen self-organizing map (KSOM) and four neural network paradigms...
One of key points in developing support vector machine (SVM) is the incorporating prior knowledge of learning task into SVM. A very common type of a prior knowledge is invariance of the input data. The research on incorporating method of invariance and SVM is an important focus for SVM in recent years, and it can help to improve the generalization performance efficiently. This paper describes and...
With the rapid growth of knowledge resources and increasing individual learning demands of users, it has been becoming more and more important to establish a kind of more comprehensible, more available and more application-oriented knowledge learning mechanism. By studying the critical factors in knowledge learning, the paper firstly proposed a novel concept of extended topic map for knowledge learning,...
A novel selective combination, optimal-weight selective ensemble (OPSEN) algorithm, is provided for the ensemble in regression tasks. It adopts the selective strategy with optimal weight matrix, whose column is the best vector corresponding to a certain training sample and can calculate the output as close to the sample target as possible. Experiment results show OPSEN is quite effective for regressor...
Process Neural Network (PNN) has an important significance in solving industry modeling problems which are related to time, but long time is cost on high dimension inputs nonlinear modeling problems. A new Improved Process Neural Networks based on KPCA and Walsh (IPNN-KPW) are proposed in this paper. KPCA method and discrete Walsh transform are used to reduce process neural network's time cost. Momentum...
Content-based image retrieval (CBIR) solutions with regular Euclidean metric usually cannot achieve satisfactory performance due to the semantic gap. Hence, relevance feedback has been adopted as a promising approach to improve the search performance. In this paper, we propose a novel idea of learning with historical relevance feedback log-data, and adopt a new methodology called "Collaborative...
In order to coordinate with and promote the scientific process of the national fencing team, we developed the decision support system for training. In fencing training, we established a two-way reasoning model based on Bayesian network and found the relationship between training process and physiological indicators. Combined with experienced knowledge and sample data, we did research on knowledge...
Models for fault diagnosis can help reduce the time taken to accurately identify faults, but the complexity of modern enterprise systems means that the process of manually model-building is itself very time-consuming. We study here the relevance of bootstrapping a diagnostic model that can then be manually refined and augmented by domain experts. We present an approach to model construction, developed...
This paper gives a deep investigation into AdaBoost algorithm, which is used to boost the performance of any given learning algorithm. Within AdaBoost, weak learners are crucial and primitive parts of the algorithm. Since weak learners are required to train with weights, two types of weak learners: artificial neural network weak learner and naive Bayes weak learner are designed. The results show AdaBoost...
Target intention inference is an important aspect of situation assessment. The evidence system of targets' intention inference is discussed according to the independent relationship between targets' intention and input evidence. The targets' intention probability inference model is proposed based on static Bayesian network. In order to expand the application domain and predigest the parameter learning...
In order to improve the application effect of the collaborative navigation control, this paper presents a Q-learning algorithm based on the path restriction by constructing the absolute distance between a mobile agent of the virtual environment and its destination into a status function of reinforcement learning. In comparison with late and former statuses, a shortest path usually can be achieved...
In this paper, we present a new approach which combines particle swarm optimization (PSO) with ensemble techniques to study credit risk assessment problems. In each iteration of the proposed method, PSO is used to solve feature subset selection problems and then nearest neighbor classifiers classify credit risk. Finally, all individual classification outputs are combined to generate the final aggregated...
Support vector machine is gaining popularity due to many attractive features and promising empirical performance in the fields of nonlinear and high dimensional pattern recognition. TSVM (transductive support vector machine) takes into account a particular test set and tries to minimize misclassifications of just those particular examples. PTSVM (progressive transductive support vector machine) can...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.