International Journal of Information Technology and Computer Science @ijitcs
Статьи журнала - International Journal of Information Technology and Computer Science
Все статьи: 1243
Extraction of Root Words using Morphological Analyzer for Devanagari Script
Статья научная
In India, more than 300 million people use Devanagari script for documentation. In Devanagari script, Marathi and Hindi are mainly used as primary language of Maharashtra state and national language of India respectively. As compared with English script, Devanagari script is reach of morphemes. Thus the lemmatization of Devanagari script is quite complex than that of English script. There is lack of resources for Devanagari script such as WordNet, ontology representation, parsing the keywords and their part of speech. Thus the overall task of information retrieval becomes complex and time consuming. Devanagari script document always carries suffixes which may cause problem in accurate information retrieval. We propose a method of extracting root words from Devanagari script document which can be used for information retrieval, text summarization, text categorization, ontology building etc. An attempt is made to design the Morphological Analyzer for Devanagari script. We have designed CORPUS containing more than 3000 possible stop words and suffixes for Marathi language. Morphological Analyzer can acts as a preliminary stage for developing any information retrieval application in Devanagari script. We have conducted the experiments on randomly selected Marathi documents and we found the accuracy of designed morphological analyzer is up to 96%.
Бесплатно
FBSEM: A Novel Feature-Based Stacked Ensemble Method for Sentiment Analysis
Статья научная
Sentiment analysis is the process of determining the attitude or the emotional state of a text automatically. Many algorithms are proposed for this task including ensemble methods, which have the potential to decrease error rates of the individual base learners considerably. In many machine learning tasks and especially in sentiment analysis, extracting informative features is as important as developing sophisticated classifiers. In this study, a stacked ensemble method is proposed for sentiment analysis, which systematically combines six feature extraction methods and three classifiers. The proposed method obtains cross-validation accuracies of 89.6%, 90.7% and 67.2% on large movie, Turkish movie and SemEval-2017 datasets, respectively, outperforming the other classifiers. The accuracy improvements are shown to be statistically significant at the 99% confidence level by performing a Z-test.
Бесплатно
FPGA Based High Accuracy Synchronous Acquisition Design for Binocular Vision System
Статья научная
This paper proposes a coarse-to-fine two-level synchronous data acquisition and transmission system for binocular stereo vision, which satisfies strict synchronous requirement of stereo vision. Specifically, this synchronization system design contains: coarse level synchronous based on hardware circuit design and the fine level synchronous based on hardware description language (HDL) design. The former includes the synchronization design of clock and external trigger. The latter utilizes a multi-level synchronous control strategy from field-level to pixel-level, which consists of field-synchronous acquisition of the two-channel video inputs, two-channel Ping-pong buffers switch control module, and pixel-synchronous bit-splicing and PCI transmission module. The experiments of synchronous acquisition and display demonstrate the high reliability and great performance of this synchronous system.
Бесплатно
Статья научная
This paper presents, a high speed FPGA implementation of fully digital controller for three-phase Z-Source Inverter (ZSI) with two switching strategies include simple boost control and maximum boost control. In this method total of blocks are based on proposed digital circuits only with combinational logic and using pipelining technique. Since it is vital to have a high speed and effective ZSI controller, a novel digital design for pulse width modulation control have been implemented for simple and maximum boost control of the ZSI. The proposed digit controllers have been successfully synthesized and implemented by Quartus II 9.1V and Cyclone II FPGA, to target device EP2C20F484C6. Achieved result demonstrates that the proposed method has features including reconfigurable, low-cost, high speed and also it is very accurate.
Бесплатно
Статья научная
In this paper we present a logical circuits design for approximate content matching implemented as finite state machines (FSM). As network speed increases the software based network intrusion detection and prevention systems (NIDPS) are lagging behind requirements in throughput of so called deep package inspection - the most exhaustive process of finding a pattern in package payloads. Therefore, there is a demand for hardware implementation. Approximate content matching is a special case of content finding and variations detection used by "evasion" techniques. In this research we will enhance the k-differentiate problem with "ability" to detect a generalized Levensthein edit distance i.e. transposition of two neighboring characters. The proposed designs are based on automata theory using the concept of state reduction and complexity minimization. The main objective is to present the feasibility of the hardware design and the trade-off between the simple next state and output functions of NFA and reduced number of required memory elements (flip-flops) of DFA.
Бесплатно
Facial Expression Classification Using Artificial Neural Network and K-Nearest Neighbor
Статья научная
Facial Expression is a key component in evaluating a person's feelings, intentions and characteristics. Facial Expression is an important part of human-computer interaction and has the potential to play an equal important role in human-computer interaction. The aim of this paper is bring together two areas in which are Artificial Neural Network (ANN) and K-Nearest Neighbor (K-NN) applying for facial expression classification. We propose the ANN_KNN model using ANN and K-NN classifier. ICA is used to extract facial features. The ratios feature is the input of K-NN classifier. We apply ANN_KNN model for seven basic facial expression classifications (anger, fear, surprise, sad, happy, disgust and neutral) on JAFEE database. The classifying precision 92.38% has been showed the feasibility of our proposal model.
Бесплатно
Factors Affecting Software Cost Estimation in Developing Countries
Статья научная
Cost is the main driving factor for all projects. When it is done correctly, it helps in the successful completion of the project. In this research we have discussed various factors that affect the estimation procedure. These include team structure, team culture, managerial style, project type (Core application or integrated application), client’s working environment. Accurate estimation is far difficult in developing countries where most of the organizations follow local standards. These inaccurate estimations lead to late delivery, less profit or in worst case complete failure. Software requirement gathering, development, maintenance, quality assurance and cost of poor quality are major groups responsible for overall cost in software production process. The exact proportion among them varies significantly in consecutive software releases, which is caused by many factors. The ever increasing need for the reliability of the software systems, especially mission critical applications in the public safety domain, raises the bar for the accuracy of prediction and estimation techniques. The accuracy of estimations in many areas brings about more concerns regarding techniques already used in the software industry. Widely deployed techniques, such as Wideband Delphi method, stress the engineering and technical aspects of the process of how estimates are prepared.
Бесплатно
Статья научная
Data on the web is constantly growing which may affect users’ ability to find relevant information within a reasonable time limit. Some of the factors previously studied that affect users searching behaviour are task difficulty and topic familiarity. In this paper, we consider a set of implicit feedback parameters to investigate how document difficulty and document familiarity affects users searching behaviour in a task-specific context. An experiment was conducted and data was collected from 77 undergraduate students of Computer science. Users’ implicit features and explicit ratings of document difficulty and familiarity were captured and logged through a plugin in Firefox browser. Implicit feedback parameters were correlated with user ratings for document difficulty and familiarity. The result showed no correlation between implicit feedback parameters and the rating for document familiarity. There was, however, a negative correlation between user mouse activities and document difficulty ratings. Also, the dataset of all the participants in the experiment was grouped according to task type and analysed. The result showed that their behaviour varies according to task type. Our findings provide more insight into studying the moderating factors that affect user searching behaviour.
Бесплатно
Farmland Intrusion Detection using Internet of Things and Computer Vision Techniques
Статья научная
Farmland security in Nigeria is still a major challenge and existing methods such as building brick fences around the farmland, installing electric fences, setting up deterrent plants with spikey branches or those that have displeasing scents are no longer suitable for farmland security. This paper presents an IoT based farmland intrusion detection model using sensors and computer vision techniques. Passive Infrared (PIR) sensors and camera sensors are mounted in strategic positions on the farm. The PIR sensor senses motion by the radiation of body heat and sends a message to the raspberry pi to trigger the camera to take a picture of the scene. An improved Faster Region Based Convolutional Neural Network is developed and used for object detection and One-shot learning algorithm for face recognition in the case of a person. At the end of the detection and recognition stage, details of intrusion are sent to the farm owner through text message and email notification. The raspberry pi also turns on the wade off system to divert an intruding animal away. The model achieved an improved accuracy of 92.5% compared to previous methods and effectively controlled illegal entry into a farmland.
Бесплатно
Fast Mapping Algorithm from WSDL to OWL-S
Статья научная
Recently semantic web services represent the most technology developed for machine to machine interaction. The problem of discovering and selecting the most suitable web service represents a challenge for semantic web services. In this paper performance evaluation of mapping algorithm from web services annotations (WSDL) to semantic annotations (OWL-S) based on ontology search engine is presented. During mapping process primitive type remains without change. The complex type are converted to OWL ontology by extracted them and passing to ontology search and standardization process without need of conversion into temporary ontology. The keywords extracted in the linguistic search phase and are extended using word net. The mapping algorithm and its modification are implemented in Java and evaluated by 310 files WSDL. The output results of two algorithms are identical. But the proposed modified algorithm is faster than mapping algorithm.
Бесплатно
Fault Persistency and Fault Prediction in Optimization of Software Release
Статья научная
This article serves two purposes: firstly, it presents an innovative methodology that increases the accuracy of fault prediction measurements. This method is based on the novel concept of "fault persistency", which enables to correct prediction metrics with a weighted value related to the module’s history. Secondly, it aims to develop operational processes from the aforesaid prediction metrics that may contribute to software construction and validation. It presents an example of an allocation methodology for resources used for testing purposes. The theoretical part is followed by an extensive experimental phase.
Бесплатно
Статья научная
Nowadays use of distributed systems such as internet and cloud computing is growing dramatically. Coordinator existence in these systems is crucial due to processes coordinating and consistency requirement as well. However the growth makes their election algorithm even more complicated. Too many algorithms are proposed in this area but the two most well known one are Bully and Ring. In this paper we propose a fault tolerant coordinator election algorithm in typical bidirectional ring topology which is twice as fast as Ring algorithm although far fewer messages are passing due to election. Fault tolerance technique is applied which leads the waiting time for the election reaching to zero.
Бесплатно
Feature Diminution by Using Particle Swarm Optimization for Envisaging the Heart Syndrome
Статья научная
Health Ecosystem is derisory in techniques to haul out the information from the database because of the lack of effective scrutiny tool to discern concealed relationships and trends in them. By applying the data mining techniques, precious knowledge can be excerpted from the health care system. Extracted knowledge can be applied for the accurate diagnosis of disease and proper treatment. Heart disease is a group of condition affecting the structure and functions of the heart and has many root causes. Heart disease is the leading cause of death in all over the world in recent years. Researchers have developed many data mining techniques for diagnosing heart disease. This paper proposes a technique of preprocessing the data set and using Particle Swarm Optimization (PCO) algorithm for Feature Reduction. After applying the PCO, the accuracy for prediction is tested. It is observed from the experiments, a potential result of 83% accuracy in the prediction. The performance of PCO algorithm is then compared with Ant Colony Optimization (ACO) algorithm. The experimental results show that the accuracy obtained from PCO is better than ACO. The performance measures are based on Accuracy, Sensitivity and Specificity. The other measures such as Kappa statistic, Mean Absolute Error, Root Mean Squared Error, True Positive Rate are also taken for evaluation. As future direction of this paper, a hybrid technique which combines PCO with Rough Set theory is suggested.
Бесплатно
Статья научная
Feature Selection (FS) is an important process to find the minimal subset of features from the original data by removing the redundant and irrelevant features. It aims to improve the efficiency of classification algorithms. Rough set theory (RST) is one of the effective approaches to feature selection, but it uses complete search to search for all subsets of features and dependency to evaluate these subsets. However, the complete search is expensive and may not be feasible for large data due to its high cost. Therefore, meta-heuristics algorithms, especially Nature Inspired Algorithms, have been widely used to replace the reduction part in RST. This paper develops a new algorithm for Feature Selection based on hybrid Binary Cuckoo Search and rough set theory for classification on nominal datasets. The developed algorithm is evaluated on five nominal datasets from the UCI repository, against a number of similar NIAs algorithms. The results show that our algorithm achieves better FS compared to two known NIAs in a lesser number of iterations, without significantly reducing the classification accuracy.
Бесплатно
Статья научная
Adverse drug reaction (ADR) is widely concerned for public health issue. ADRs are one of most common causes to withdraw some drugs from market. Prescription event monitoring (PEM) is an important approach to detect the adverse drug reactions. The main problem to deal with this method is how to automatically extract the medical events or side effects from high-throughput medical events, which are collected from day to day clinical practice. In this study we propose a novel concept of feature matrix to detect the ADRs. Feature matrix, which is extracted from big medical data from The Health Improvement Network (THIN) database, is created to characterize the medical events for the patients who take drugs. Feature matrix builds the foundation for the irregular and big medical data. Then feature selection methods are performed on feature matrix to detect the significant features. Finally the ADRs can be located based on the significant features. The experiments are carried out on three drugs: Atorvastatin, Alendronate, and Metoclopramide. Major side effects for each drug are detected and better performance is achieved compared to other computerized methods. The detected ADRs are based on computerized methods, further investigation is needed.
Бесплатно
Feature Selection using a Novel Particle Swarm Optimization and It’s Variants
Статья научная
Feature selection has been keen area of research in classification problem. Most of the researchers mainly concentrate on statistical measures to select the feature subset. These methods do not provide a suitable solution because the search space increases with the feature size. The FS is a very popular area for applications of population-based random techniques. This paper suggests swarm optimization technique, binary particle swarm optimization technique and its variants, to select the optimal feature subset. The main task of the BPSO is the selection of the features used by the SVM in the classification of spambase data set. The results of our experiments show a very strong relation between number of features and accuracy. Comparison of the optimized results and the un-optimized results showed that the BPSO-MS method could significantly reduce the computation cost while improving the classification accuracy.
Бесплатно
Feature Selection with Targeted Projection Pursuit
Статья научная
The selection of attributes becomes more important, but also more difficult, as the size and dimensionality of data sets grows, particularly in bioinformatics. Targeted Projection Pursuit is a dimension reduction technique previously applied to visualising high-dimensional data; here it is applied to the problem of feature selection. The technique avoids searching the powerset of possible feature combinations by using perceptron learning and attraction-repulsion algorithms to find projections that separate classes in the data. The technique is tested on a range of gene expression data sets. It is found that the classification generalisation performance of the features selected by TPP compares well with standard wrapper and filter approaches, the selection of features generalises more robustly than either, and its time efficiency scales to larger numbers of attributes better than standard searches.
Бесплатно
Feature Selection: A Practitioner View
Статья научная
Feature selection is one of the most important preprocessing steps in data mining and knowledge Engineering. In this short review paper, apart from a brief taxonomy of current feature selection methods, we review feature selection methods that are being used in practice. Subsequently we produce a near comprehensive list of problems that have been solved using feature selection across technical and commercial domain. This can serve as a valuable tool to practitioners across industry and academia. We also present empirical results of filter based methods on various datasets. The empirical study covers task of classification, regression, text classification and clustering respectively. We also compare filter based ranking methods using rank correlation.
Бесплатно
Feature selection to classify healthcare data using wrapper method with pso search
Статья научная
As a result of the rapid development of technology, data that contain a large number of features are produced from various applications such as biomedical, social media, face recognition, etc. Processing of these data is a challenging task to existing data mining and machine learning algorithms to make the decision. To reduce the size of the data for processing, a feature selection technique is needed. The feature selection is a well-known attribute selection or variable selection. The objective of the feature selection is to minimize the number of attributes contains in the dataset by eliminating the unwanted and repeated attributes to improve the classification accuracy and reduce the computation cost. Although various feature selection methods are proposed, in literature, to classify the healthcare data especially cancer diagnosis, finding an informative feature for medical datasets has still remained a challenging issue in the data mining and machine learning domain. Therefore, this paper presents a feature selection approach with the wrapper method (WFS) using particle swarm optimization (PSO) search to improve the accuracy of healthcare data classification. This work is evaluated on five benchmark medical datasets publicly available from the UCI machine learning repository. The experimental results showed that the WFS-PSO approach produces higher classification accuracy applied to different classification algorithms.
Бесплатно
Fingerprints, Iris and DNA Features based Multimodal Systems: A Review
Статья научная
Biometric systems are alternates to the traditional identification systems. This paper provides an overview of single feature and multiple features based biometric systems, including the performance of physiological characteristics (such as fingerprint, hand geometry, head recognition, iris, retina, face recognition, DNA recognition, palm prints, heartbeat, finger veins, palates etc) and behavioral characteristics (such as body language, facial expression, signature verification, speech recognition, Gait Signature etc.). The fingerprints, iris image, and DNA features based multimodal systems and their performances are analyzed in terms of security, reliability, accuracy, and long-term stability. The strengths and weaknesses of various multiple features based biometric approaches published so far are analyzed. The directions of future research work for robust personal identification is outlined.
Бесплатно
