An integrated approach of mining association rules and meta-rules based on a hyper-structure is put forward. In this approach, time serial databases are partitioned according to time segments, and the total number of ...
详细信息
ISBN:
(纸本)3540269231
An integrated approach of mining association rules and meta-rules based on a hyper-structure is put forward. In this approach, time serial databases are partitioned according to time segments, and the total number of scanning database is only twice. In the first time, a set of 1-frequent itemsets and its projection database are formed at every partition. then every projected database is scanned to construct a hyper-structure. through mining the hyper-structure, various rules, for example, global association rules, meta-rules, stable association rules and trend rules etc. can be obtained. Compared with existing algorithms for mining association rule, our approach can mine and obtain more useful rules. Compared with existing algorithms for meta-mining or change mining, our approach has higher efficiency. the experimental results show that our approach is very promising.
the proceedings contain 60 papers. the topics discussed include: predicting software suitability using a Bayesian belief network;parallel algorithm for control chart patternrecognition;data-centric automated data min...
详细信息
ISBN:
(纸本)0769524958
the proceedings contain 60 papers. the topics discussed include: predicting software suitability using a Bayesian belief network;parallel algorithm for control chart patternrecognition;data-centric automated data mining;a Bayesian kernel for the prediction of neuron properties from binary gene profiles;new filter-based feature selection criteria for identifying differentially expressed genes;a new clustering algorithm using message passing and its applications in analyzing microarray data;iterative weighting of phylogenetic profiles increases classification accuracy;integrating knowledge-driven and data-driven approaches for the derivation of clinical prediction rules;sparse classifiers for automated heart wall motion abnormality detection;segmenting brain tumors using alignment-based features;and the application of machinelearning techniques to the prediction of erectile dysfunction.
Derivative free optimization methods have recently gained a lot of attractions for neural learning. the curse of dimensionality for the neural learning problem makes local optimization methods very attractive;however ...
详细信息
ISBN:
(纸本)3540269231
Derivative free optimization methods have recently gained a lot of attractions for neural learning. the curse of dimensionality for the neural learning problem makes local optimization methods very attractive;however the error surface contains many local minima. Discrete gradient method is a special case of derivative free methods based on bundle methods and has the ability to jump over many local minima. there are two types of problems that are associated withthis when local optimization methods are used for neural learning. the first type of problems is initial sensitivity dependence problem - that is commonly solved by using a hybrid model. Our early research has shown that discrete gradient method combining with other global methods such as evolutionary algorithm makes them even more attractive. these types of hybrid models have been studied by other researchers also. Another less mentioned problem is the problem of large weight values for the synaptic connections of the network. Large synaptic weight values often lead to the problem of paralysis and convergence problem especially when a hybrid model is used for fine tuning the learning task. In this paper we study and analyse the effect of different regularization parameters for our objective function to restrict the weight values without compromising the classification accuracy.
We estimate the speed of texture change by measuring the spread of texture vectors in their feature space. this method allows us to robustly detect even very slow moving objects. By learning a normal amount of texture...
详细信息
this paper proposes a new algorithm to improve learning performance in support vector machine by using the Kernel Relaxation and the dynamic momentum. Compared withthe static momentum, the dynamic momentum is simulta...
详细信息
this paper extends the idea of weighted distance functions to kernels and support vector machines. Here, we focus on applications that rely on sliding a window over a sequence of string data. For this type of problems...
详细信息
ISBN:
(纸本)0780390911
this paper extends the idea of weighted distance functions to kernels and support vector machines. Here, we focus on applications that rely on sliding a window over a sequence of string data. For this type of problems it is argued that a symbolic, context-based representation of the data should be preferred over a continuous, real format as this is a much more intuitive setting for working with (weighted) distance functions. It is shown how a weighted string distance can be decomposed and subsequently used in different kernel functions and how these kernel functions correspond to inner products between real vectors. As a case-study named entity recognition is used with information gain ratio as a weighting scheme.
this paper discusses a consistency in patterns of language use across domain-specific collections of text. We present a method for the automatic identification of domain-specific keywords - specialist terms - based on...
详细信息
Steering an autonomous vehicle requires the permanent adaptation of behavior in relation to the various situations the vehicle is in. this paper describes a research which implements such adaptation and optimization b...
详细信息
Imitation has been regarded as one of the key technologies indispensable for communication since mirror neuron [1] made a big sensation not only in physiology but also in other disciplines such as cognitive science, a...
详细信息
ISBN:
(纸本)0780392256
Imitation has been regarded as one of the key technologies indispensable for communication since mirror neuron [1] made a big sensation not only in physiology but also in other disciplines such as cognitive science, and even robotics as well. Unlike a simple copy of human motion trajectories, imitation may include more important role of human motion recognition. that is, observing other's behavior may recall the self motion through the mirror system, and this might be considered as the key component of recognition, communication and even language acquisition [2]. It is an interesting question;in what point imitation faculty is effective for communication learning? if a robot can imitate normal motions of a human partner, is it easy to read mind of the partner for the robot? this paper is the first step to those questions. In this paper, we aim at building a human-robot communication system and propose an observation-to- motion mapping system as the first step towards the final goal, learning natural communication. this system enables a humanoid platform to imitate the observed human motion, that is, a mapping from observed human motion data to its own motor commands. To realize this capability, we suppose a human partner who kindly imitates the robot motion, and the system associates bothdata of the robot somatosensory information (the set of joint angles) and observed human motions imitated from the robot motions, each of which is self-organized onto two dimensional maps using the isometric feature mapping (ISOMAP) algorithm [3] for data reduction, respectively, beforehand. A neural network is utilized for this association based on which the humanoid can imitate human motions. this system is applied to interaction rule learning with a human partner who knows the rule and reacts to the humanoid action according to them. On the other hand, the humanoid does not know the rule at the beginning but gradually learns the rule by using its own reaction rule: to just imita
暂无评论