Earticle

Semantic similarity has attracted great concern for a long time in artificial intelligence, psychology and cognitive science. In recent years the measures based on WordNet have shown its talents and attracted great concern. Many measures have been proposed. The paper contains a review of the state of art measures, including path based measures, information based measures, feature based measures and hybrid measures. The features, performance, advantages, disadvantages and related issues of different measures are discussed. Finally the area of future research is described..

Fuel ratio (FR) is the mass ratio of air and fuel trapped inside a cylinder before combustion begins, and it affects engine emissions, fuel economy, and other performances. For a dual fuel engine equipped with both port-fuel-injection (PFI) and direct injection (DI) systems, the fuel ratio is the ratio of the first fuel and total fuel masses. In this research, a multi-input-multi-output baseline chattering free sliding mode methodology scheme is developed with guaranteed stability to simultaneously control fuel ratios to desired levels under various air flow disturbances by regulating the mass flow rates of engine PFI and DI injection systems. A baseline estimator with varying parameter gain is designed with guaranteed stability to allow implementation of the proposed state feedback sliding mode methodology into a MATLAB simulation environment, where the sliding mode strategy is implemented into a model engine control module (“software”). The baseline sliding methodology performance was compared with a well-tuned baseline multi-loop PID controller through MATLAB simulations and showed improvements, where MATLAB simulations were conducted to validate the feasibility of utilizing the developed controller and state estimator for automotive engines. A baseline sliding mode tracking methodology is developed to regulate the transient and steady state FR based upon a control oriented model of the engine PFI wall wetting dynamics. The proposed tracking method is designed to optimally track the desired FR by minimizing the error between the trapped in-cylinder mass and the product of the desired FR and fuel mass over a given time interval. The performance of the optimal proposed tracking methodology was compared with the conventional fueling control based on the inverse fueling dynamics through simulations and showed improvement over the baseline conventional inverse fueling dynamics methodology.

To meet the demands of wireless sensor networks (WSNs) where data are usually aggregated at a single source prior to transmitting to any distant user, there is a need to establish a tree structure inside to aggregate data. In this paper, an adaptive energy aware data aggregation tree (AEDT) is proposed. The proposed tree uses the maximum energy available node as the data aggregator node. The tree incorporates sleep and awake technology where the communicating node and the parent node are only in awake state rest all the nodes go to sleep state saving the network energy and enhancing the network lifetime. When the traffic load crosses the threshold value, then the packets are accepted adaptively according to the communication capacity of the parent node. The proposed tree maintains a memory table which stores the value of each selected path. Path selection is based on shortest path algorithm where the node with highest available energy is always selected as forwarding node. By simulation results, we show that our proposed tree enhances network lifetime minimizes energy consumption and achieves good delivery ratio with reduced delay.

In terms of machine learning theory, the intrinsic geometrical structure of the original data space is usually embedded in the low-dimensional manifold. The extraction of optimized manifold features could improve the performance of clustering. This paper presents a new spectral clustering method called local topology preserving indexing (LTPI). In this algorithm, the data are projected into a low-dimensional feature space in which the distances between the data points in the same local patches are minimized and the distances from the data points outside these patches are maximized simultaneously. The proposed LTPI method can effectively discover the intrinsic local topologies embedded in original high-dimensional space. The comparison experiments for document clustering demonstrate its effectiveness.

Email has become one of the fastest and most economical forms of communication. However, the increase of email users has resulted in the dramatic increase of spam emails during the past few years. As spammers always try to find a way to evade existing filters, new filters need to be developed to catch spam. Generally, the main tool for email filtering is based on text classification. A classifier then is a system that classifies incoming messages as spam or legitimate (ham) using classification methods. The most important methods of classification utilize machine learning techniques. There are a plethora of options when it comes to deciding how to add a machine learning component to a python email classification. This article describes an approach for spam filtering using Python where the interesting spam or ham words (spam-ham lexicon) are filtered first from the training dataset and then this lexicon is used to generate the training and testing tables that are used by variety of data mining algorithms. Our experimentation using one dataset reveals the affectivity of the Naïve Bayes and the SVM classifiers for spam filtering.

Grey system theory and rough set theory are two different mathematical tools that are used to deal with uncertain or incomplete information, and yet they are relevant and complementary to a certain degree. The appropriate hybrid of the two theories can overcome the shortages of their definitions and applications and thus has more powerful functions. This paper proposes An Integrated Methodology that extracting decision rules based on combining grey system and rough set theory. The effectiveness of the proposed methodology was verified by application of this methodology to discover grade rules of electrical transformer evaluation.

This paper analyzes the validity of temperature maps obtained by means of single and mixed interpolation methods. In this context, several interpolation methods are used for the temperature mapping: inverse distance square (IDS), ordinary kriging (OK) and co-kriging (CK) and mixed methods (combined global, local and geostatistical methods). And the validity of the maps is checked through the cross-validation error statistics presented in terms of Mean absolute error (MAE) and root-mean-square error (RMSIE). The results show that the spatial temperature distribution formed by IDS, OK and CK show similarity in general while the CK method better reflects the temperature distribution. Moreover, the best result for temperature mapping is obtained using the mixed method, which suggests that the correction of regression models using residuals improves the interpolation accuracy. Reliable calculation for agricultural management and improvement of climatic models at local scales can be obtained with increased efficiency.

Conception and calculation method of information entropy is customized for Sudoku puzzles and a corresponding algorithm is designed to solve Sudoku puzzles. The definitions of inverse information entropy and information amount for inverse information entropy are also introduced and directly used instead of information entropy in order to simplify the solving procedure. Experimental results show that the algorithm has better time efficiency than available methods including generic algorithms and rule based algorithms and it can solve not only unique-solution puzzles (including extremely difficult puzzles) but also multiple-solution puzzles. It is concluded that information entropy can be used for grading Sudoku puzzles as well.

With the advance in the electronic devices, digital publications become more and more popular in our life. However, the properties of digital contents make themselves be easily copied and transferred if there is not any proper protection for them. Hence, it is a critical issue for publication provider to eectively control and distribute their digital publications. Digital Rights Management is a mechanism, which might congregate various techniques to protect the rights of digital publication from copyrights violations. Moreover, Wrapper-based Digital Rights Management technique applies encapsulating digital contents by packaging content and monitoring by API-Hook to control and protect them, which provide a way to authenticate users by users' machine serial number or smart card via network. Hence, users may use the digital contents without changing their digital content player. According to the denition of Digital Rights Management, this paper provides a digital publication issuing mechanism, which supports superdistribution for advertising digital publications eectively and improving development of digital contents.

Many traditional time series model has been widely applied in forecasting Problem. However, the previous time series methods still have some constraints: (1) conventional time series models only considered single variable; (2) traditional fuzzy time series model determined the interval length of linguistic value subjectively; (3) selecting variables depended on personal experience and opinion. Hence, this paper proposes a novel hybrid fuzzy time series model based on fitting function to forecast TAIEX (Taiwan Stock Exchange Capitalization Weighted Stock index). The proposed model employed Pearson’s correlation to select important technical indicators objectively, and the proposed model utilized fitting function to forecast TAIEX Index. In verification, the collected TAIEX datasets from 1998/01/03 to 2002/12/31 are used as experimental dataset and the root mean square error (RMSE) as evaluation criterion. The results show that the proposed model outperforms the listing models in accuracy.

Hyper-spectral Atmospheric Infrared Sounder has many channels. Although, it has been carried out channel selection in the actual application process, but compared with the previous detector, the channel also seems more and thus easy to weak correlation between channels, which is not obey classical variational assimilation theory requires observation error is not related. Hyper-spectral sounder covering CO2 and water vapor absorption band channels, the nonlinear of water vapor channel brightness temperature is strong, the practical application of these channel combinations into the assimilation system together, the brightness temperature information mutual influence and feedback among channels are more complex. Different from the classical variational assimilation that given observation error is not change later, using the observational error re-estimation and robust variational assimilation based on Huber function to study the hyper-spectral sounder brightness temperature assimilation. The ideal experiments show that the observation error re-estimates and robust variational method assimilate hyper-spectral atmospheric infrared sounder can get better assimilation effect than the classical one.