摘要 :
We aim to find the effect of diversity on combiner performance. Three diversity measures are used to calculate the diversity of combined classifiers. We aim to identify which measure is closely related to the combiner performance....
展开
We aim to find the effect of diversity on combiner performance. Three diversity measures are used to calculate the diversity of combined classifiers. We aim to identify which measure is closely related to the combiner performance. Three combiner types are used; Bagging and a conventional three classifier system, in which three classifier types are used; backpropagation neural network, bayesian and k-nearest neighbor classifiers. Additionally we experiment with a feature based combiner system proposed by Alkoot [13,14]. Results obtained on real data indicate the diversity measure to be higher for systems with higher classification rate, if it outperforms other classifiers by a large margin. Otherwise, if the performances of the compared systems are close the diversity measure may not be higher for the best system. On many occasions the diversity measures were not good indicators of system performance. On some instances we found that the more diverse system did not yield a better performance.
收起
摘要 :
In this paper, a new automated information retrieval system is presented. The design of such system relies on the components used in computer based information system (CBIS). Its efficiency and effectiveness play an important role...
展开
In this paper, a new automated information retrieval system is presented. The design of such system relies on the components used in computer based information system (CBIS). Its efficiency and effectiveness play an important role in the success of the automated information retrieval system. The main objective of this work is to help Internet users to find the required information with high performance. Finding information on an object's visual features is useful when specific keywords for the object are not known. Since intelligent mobile agent technology is expected to be a promising technology for information retrieval, there is a number of intelligent mobile agent based-information retrieval approaches have been proposed in recent years. Multiple information agents continuously traverse the Internet and collect images that are subsequently indexed based on image information such as the URL location, size, type and the date of indexation. In the search phase, the intelligent mobile agent receives the image of object as a query. Then, the fast searching algorithm presented in [31] is used to search the set of web pages that contain information about the object. Furthermore, different models of information retrieval are studied. In addition, intelligent adversarial information retrieval is described. Moreover, different methods for measuring the performance of information retrieval are discussed.
收起
摘要 :
It is common in real life to find fuzzy information that comes from subjective judgments or the imprecision in measured data. Fuzzy approaches have been used to extend database systems in storing and updating imprecise information...
展开
It is common in real life to find fuzzy information that comes from subjective judgments or the imprecision in measured data. Fuzzy approaches have been used to extend database systems in storing and updating imprecise information (data) and in processing imprecise queries. Consider a fuzzy query: find name, grade of quite good students and just tall students where age > 15. This query includes two fuzzy concepts: good student and tall student and one crisp query criteria (i.e. age > 15). In this paper we present a formalization to process natural language fuzzy (expressive) queries and to return fuzzy results for crisp query criteria. Our formalization is general that can be particularized for implementation in variety of database platforms i.e. fuzzy web search, information systems supporting fuzzy data etc. Our approach only makes the fuzzy query writing much simpler and easier than conventional query writing but also close to human like thinking due to its true fuzzy nature. We also provide an operational semantics for fuzzy query processing which can be followed for multiple data types i.e. numeric, text, graphics etc. Our approach supports fuzzy querying for not only fuzzy data but also for missing data; hence enabling us to get query results closer to human thinking and expectations. It is an expressive model that let to make human-like (i.e. fuzzy) consults.
收起
摘要 :
The purpose of this research was to study the status, needs and problems of using online Course in learning of Industrial Technology at Rajamangala University of Technology Isan. The sampling groups were 462 instructors and 1170 s...
展开
The purpose of this research was to study the status, needs and problems of using online Course in learning of Industrial Technology at Rajamangala University of Technology Isan. The sampling groups were 462 instructors and 1170 students from stratified sampling. A research tool used was a questionnaire. The statistics used in data analysis were percentage, standard deviation and one-way ANOVA. The results of this research revealed that most instructors (84%) had never designed the online course and did not have the knowledge of online course design. The learners (82%) never studied online course. According to the Needs showed that the increasing of speed networks, knowledge of online course construct, using online course method, learning towards the online course and the amount of computers were at a high level. In the other hand, the Problems of instructors showed that non policy and supporting in using of online course from campus or universities, the lack of budget in equipments buying, more teaching load and another works to do and no times to construct online course which were at a high level. The result of comparison between the Needs and Problems revealed that the construction and using online course of the instructors which were in difference campus, the instructors who were in Khon Kaen Campus, Northeastern Campus and Sakonnakhon Campus had the needs in continuing online course training had difference significant at a level of .05. The result of comparison between the Needs and Problems on using and learning towards the online course revealed that the instructors who were in Khon Kaen Campus, Northeastern Campus and Sakonnakhon Campus had the needs on extension opening-period in continuing online course had difference significant at a level of .05.
收起
摘要 :
With a constant increase of servers worldwide, estimated in 2010 to be "50 million servers in the world today", Napier, A. L. [1]; the power needed to run server farms being "over 1% of the world-wide electricity consumption", Fet...
展开
With a constant increase of servers worldwide, estimated in 2010 to be "50 million servers in the world today", Napier, A. L. [1]; the power needed to run server farms being "over 1% of the world-wide electricity consumption", Fettweis and Zimmermann [2]. This is inevitably coupled with more heat dissipation, leading to a cooling problem that constitutes 200% of the direct power consumption in server farms, Schott [3]. Considering the fact that "most servers are running at 5-15% of their capacity" Siebert [4]; many worldwide developments in technologies and methodologies were directed towards reducing power consumption in server farms; rather than tackling the most imperative problem of under utilization. The mathematical model presented in this research aims at reducing the power consumption by minimizing the number servers (and ancillary equipment), that need to be on, while meeting the required demand. The model guarantees arriving at the minimal operating power. Applying the proposed approach to three formulated examples resulted in reducing the percentage of idle servers from 7.3 %, to 2.1 % and then to 0% idle servers, respectively.
收起
摘要 :
In previous work [23,24], a fast systematic method for minimization of the Boolean functions was presented. Such method is a simple because there is no need for any visual representation such as Karnough map or arrangement techniq...
展开
In previous work [23,24], a fast systematic method for minimization of the Boolean functions was presented. Such method is a simple because there is no need for any visual representation such as Karnough map or arrangement technique such as Tabulation method. Furthermore, it is very easy and fast for programming. In addition, it is suitable for boolean function with large number of variables (more than 4 variable). Moreover, it is very simple to understand and use. In this paper, the simplified functions are implemented with minimum amount of components. A powerful solution for realization of complex functions is given. This is done by using modular neural nets (MNNs) that divide the input space into several homogenous regions. Such approach is applied to implement XOR functions, 16 logic function on one bit level, and 2-bit digital multiplier. Compared to previous non- modular designs, a clear reduction in the order of computations and hardware requirements is achieved.
收起
摘要 :
The present study tried to cralify that the storyteller controlled the time length of pause at the final period in each sentence to support children's comprehension and suggest the way in which given findings are applied to readin...
展开
The present study tried to cralify that the storyteller controlled the time length of pause at the final period in each sentence to support children's comprehension and suggest the way in which given findings are applied to reading the text by the use of synthesized speech. During listening to the sound of telling a story text, a listener constructs a mental microworld that is refer to as the term situation model in the discourse psychology. Unlike the storyteller who can read the text at his own pace, however, the listener's construction of the situation model should be influenced by the way in which a storyteller reads the text. Because the event-indexing model indicated that comprehending time of a sentence in a story is progressively longer to the extent that there are continuity breaks on more dimensions (e.g., space, time, characters and objects), we focused on the relationship between the number of mental operations to comprehend a sentence and the time length of pause at the final period in each sentence of the story read by the storyteller. In order to estimate the mental operations that are required to construct the situation model, we devised the multilayered frame representation. Then, we defined a sequence of the operations to transform the case frame representation of each sentence to the multilayered frame representation and examined the relationship between it and the time length of pause at period in the speech sound of a vastly-experienced storyteller. The results indicated that the storyteller controlled the time length of pause according to the number of the operations so as to help the listener to comprehend the story.
收起
摘要 :
Human computer Interaction based on emotional modelling and physical views, collectively; has been investigated and reported in this paper. Two types of ontology have been presented to formalize a patient state: mental ontology re...
展开
Human computer Interaction based on emotional modelling and physical views, collectively; has been investigated and reported in this paper. Two types of ontology have been presented to formalize a patient state: mental ontology reflecting the patient mental behavior due to certain disorder and physical ontology reflecting the observed physical behavior exhibited through disorder. These two types of ontology have been mapped and aligned for reasoning using a simple Bayesian Network for causal reasoning to define what we call as simple case diagnosis. We have constructed an integrated computerized model which reflects a human diagnostician as computer model and through it; an integrated interaction between that model and the real human user (patient) is utilized for 1st stage diagnosis purposes.
收起
摘要 :
Brain tumor diagnosis is a very crucial task. This system provides an efficient and fast way for diagnosis of the brain tumor. Proposed system consists of multiple phases. First phase consists of texture feature extraction from br...
展开
Brain tumor diagnosis is a very crucial task. This system provides an efficient and fast way for diagnosis of the brain tumor. Proposed system consists of multiple phases. First phase consists of texture feature extraction from brain MR images. Second phase classify brain images on the bases of these texture feature using ensemble base classifier. After classification tumor region is extracted from those images which are classified as malignant using two-stage segmentation process. Segmentation consists of skull removal and tumor extraction phases. Quantitative results show that our proposed system performed very efficiently and accurately. We achieved accuracy of classification beyond 99%. Segmentation results also show that brain tumor region is extracted quite accurately.
收起
摘要 :
Decommitment of agents has been studied in the context of self-interested agent societies, primarily using a leveled commitment protocol in which agents committing a breach of contract are penalized in various ways. Limited resear...
展开
Decommitment of agents has been studied in the context of self-interested agent societies, primarily using a leveled commitment protocol in which agents committing a breach of contract are penalized in various ways. Limited research has investigated decommitment in collaborative agent populations. This research investigates decommitment in a collaborative agent society, in which agents make decisions and act in a manner which maximizes the global benefit. In this environment, decommitment takes on a different nuance, in that collaborative agent behavior can be expected not to cause harm to the agent society as a whole, even in decisions not to honor prior commitments. The operating environment in this study is both time-bounded and resource constrained. This research studies three levels of agent decommitment: 1) no decommitment, 2) unilateral decommitment, and 3) negotiated decommitment. Results indicate that negotiated decommitment generally performed as hypothesized, that is, overall performance was significantly improved over the baseline condition, and, with certain criteria, its performance was significantly improved over the unilateral decommitment condition.
收起