keivan borna

Since the discovery and representation of subjects contained in information resources is the most important goal in information description, analysis and retrieval systems, and since subject search is the most common type of user search in databases and library catalogs, in this study what is "subject" and how to analyze and extract subject matters in the process of determining subject keywords for an information resource, as well as some research and operational perspectives are discussed.

This conceptual paper uses documentary research method to examine the concepts of "subject" and "subject analysis", presented in a selection of related research works in the field of Information Science, as well as considering some corresponding concepts in the field of Computer Science and Natural Language Processing.

Despite various interpretations of what "subject" is, which some assume to be an axiom, and its definition unnecessary, there is consensus, explicitly or implicitly, on its meaning of "aboutness". Subject analysis of information resources, as part of the indexing process, is the analysis and identification of stated topics and concepts and/or obvious features of the information source that may involve manpower, computer algorithms designed to identify textual terms or the combination of the two is done. Each of these methods has its strengths and weaknesses. Search and retrieval problems due to the multiplicity of subject analysis tools and keyword assignment, low usage rate of standard descriptive schemas, inconsistencies between indexed terms assigned by a single indexer at different times, as well as between multiple indexers for the same information resource, human error in general, the inconsistency of users’ search terms with the assigned keywords, and the lack of on time description of resources due to the growing production of them are among the weaknesses of the human subject analysis approach. The application of automated methods of artificial intelligence and natural language processing offers promising prospects for increasing speed and consistency in various processes of describing and organizing information, including the extraction of subject keywords. However, evaluating the efficiency of the output of these methods, alone or in comparison with the keywords of human production, specially from the perspective of users, is required.

In the competition of machine algorithms with the human mind in the analysis and recognition of resource subjects, the human mind excels; whether it is an indexer who expresses in his own language what the source is about, or a designer who can design a machine that mimics the computational steps of his mind to saves time and resources. Libraries and information centers can, far from rushing to replace conventional processes and procedures, use the various methods and tools available from the Natural Language Processing and Machine Learning fields to design and develop automated systems for indexing and extracting or assigning subject keywords - and in the larger perspective, automated classification- to use resources (financial, human, and time) for the ultimate goal of maximizing productivity, shortening the subject access path to resources for users, and facilitating the description process.

The Internet of Things provides instant access to information about the physical world and the objects within it, leading to new services and increasing efficiency and productivity. The wireless sensor network is an important network infrastructure in the Industrial Internet of Things and user authentication is used as a basic security mechanism to authenticate users to wireless sensor networks. In this article, we intend to provide a new way to improve the security of authentication using image processing. The analysis was performed by MATLAB software. The results presented in this article, which includes seven steps, have a 93% correct detection rate in pupil identification. These seven steps are: 1. Noise reduction, 2. Finding the outer border of the pupil, 3. Separating the eyelashes, 4. Finding the border of the eyelids, 5. Finding the outer border of the iris, 6. Separating the iris area and 7. Extracting the feature and encoding the pixels by the elliptic curve cryptography (ECC) method. With the measurements and experiments performed, it was found that the proposed method in identifying the eyelid border by the quadratic equation method is more efficient and faster in terms of time than the third-degree equation and the Huff parabolic conversion method. In order to extract the feature, the effective parameters in the SAIF feature extraction algorithm were examined and measured, and the optimal parameters were selected. The Sigma parameter with a value of 2.5 and the Octave parameter with a value of 4 should be considered as the best values. Also, in order to evaluate the resistance of the proposed method to error factors such as the angle, brightness and scale, the proposed method was tested and it was proved that the method has the appropriate resistance resolution in different conditions.

Deciding whether a musical rhythm is good or not, depends on many factors like geographical conditions of a region, culture, the mood of society, the view of rhythm over years, and so on. In this paper, we want to make a decision from the scientific point of view, using geometric features of rhythms, about bad ones. The researchers who are investigating the relationship between geometry and music, certainly realize that there is a big vacuum in this regard, not using computers to detect a good or bad rhythm. Here, using computer programming and applying geometric features to more than four thousand rhythms, we decide on the bad musical rhythms. Then we present algorithms for deciding about bad rhythms using geometrical features.

Spam is an unwanted email that is harmful to communications around the world. Spam leads to a growing problem in a personal email, so it would be essential to detect it. Machine learning is very useful to solve this problem as it shows good results in order to learn all the requisite patterns for classification due to its adaptive existence. Nonetheless, in spam detection, there are a large number of features to attend as they play an essential role in detection efficiency. In this article, we're working on a feature selection method to e-mail spam. This approach is considered a hybrid of optimization algorithms and classifiers in machine learning. Binary Whale Optimization (BWO) and Binary Grey Wolf Optimization (BGWO) algorithms are used for feature selection and K-Nearest Neighbor (KNN) and Fuzzy K-Nearest Neighbor (FKNN) algorithms are applied as the classifiers in this research. The proposed method is tested on the "SPAMBASE" datasets from UCI Machine learning Repesotries and the experimental results revealed the highest accuracy of 97.61% on this dataset. The obtained results indicateed that the proposed method is suitable and capable to provide excellent performance in comparison with other methods.

Market forecasting, like stock's market, with high volume of transactions have an effect on researchers and investors and get their attention. Important issue factors in any investment decision are risk and turnover. Understanding market momentum gives the ability to predict future movements. The ability to predict in a market economy, enables to achieve a higher turnover by reducing risk and avoiding financial losses. News plays an important role in the process of evaluating the current stock price. The development of data mining methods, computational intelligence and machine learning algorithms have led to new models of prediction. phpCrawler is a php base content crawler that use of DomCralwer and Guzzle packages for storing web data. With this tool, the news releases are stored and categorized from 17 News Agency.    Then, by using text mining and Support Vector Machine with different kernel, predict stock price direction. In this research use 948990, news has been stored from 17 news agencies. More than 300,000 news regarding political and economic categories were used and stock prices of chemicals between November 2017 till March 2018 (123 trading days) were studied. The results show that by using the linear kernel Support Vector Machine algorithm, the prediction accuracy of average price movement reached to 83%. Using nonlinear kernel Support Vector Machine with poly kernel increased two percent prediction accuracy to 85% on average and other kernel had poorer prediction.

Convex hull of some given points is the intersection of all convex sets containing them. It is used as primary structure in many other problems in computational geometry and other areas like image processing, model identification, geographical data systems, and triangular computation of a set of points and so on. Computing the convex hull of a set of point is one of the most fundamental and important problems of computational geometry. In this paper a new algorithm is presented for computing the convex hull of a set of random points in the plane by using a sweep-line strategy. The sweep-line is a horizontal line that is moved from top to bottom on a map of points. Our algorithm is optimal and has time complexity $O(nlogn)$ where $n$ is the size of input.

The internet of things will change our life in future significantly and will make the impossible, possible. A large volume of big data which is produced or taken by Internet of Things (IOT) contains valuable and useful information. By the prevalence of the wireless apparatuses technology such as Bluetooth, detection of radio frequency (RFID), Wi-Fi and data services on telephone, sensor, actuators and nodes embedded in the equipment, Wireless Sensor Networks (WSN) the internet of things has already passed its primary stages and is in the threshold of changing the current static internet into a fully integrated internet. Data mining; too, with no doubt plays a large role in smartness of the system and subsequently, provides suitable services and environment in offering services. Also, data mining techniques are used to cluster nodes and determine cluster head, in wireless sensor networks. This paper introduces the internet of things, architecture and its applications.

One of the most common problems with computer networks is the amount of information in these networks. Meanwhile searching and getting inform about content of textual document, as the most widespread forms of information on such networks, is difficult and sometimes impossible. The goal of multi-document textual summarization is to produce a pre-defined length summary from input textual documents while maximizing documents’ content coverage. This paper presents a new approach for textual document summarization based on paraphrasing and textual entailment relations and formulating the problem as an optimization problem. In this approach the sentences of input documents are clustered according to paraphrasing relation and then the entailment score and final score of a fraction of the header sentences of clusters which have the best score according to the user query is calculated. Finally, the optimization problem is solved via greedy and dynamic programming approaches and while selecting the best sentences, the final summary is generated. The results of implementing the proposed system on standard datasets and evaluation via ROUGE system show that the proposed system outperforms the state-of-the-art systems at least by 2.5% in average.

Automatic identification and classification of the Iranian traditional music scales (Dastgāh) and melody models (Gusheh) has attracted the attention of the researchers for more than a decade. The current research aims to review conducted researches on this area and consider its different approached and obstacles. The research approach is content analysis and data collection method is based on the documentary-library study. Findings indicated that the main obstacles and reasons for the inefficiency of this area researches are due to performing the researches in parallel and individually, lack of a coherent dataset for Iranian traditional music and also, lack of researcher’s knowledge of the theoretical foundations of Iranian traditional music. It could be considered two main approaches for researches in automatic identification of the Iranian traditional music. The major researches has been published until now, is conducted based on Macro approach, which merely considers the scales of five Dastgāhs in distinguishing them from each other. Since Dastgāhs division does not have enough authenticity and there is no consensus on the number of Dastgāhs and their boundaries among the Iranian music theorists, Micro approach has been suggested for future researches, which tries to identify short melody models (Gushehs) based of melody sequences of representative phrase of the Gusheh.

Predicting the price of electricity is very important because electricity can not be stored. To this end, parallel methods and adaptive regression have been used in the past. But because dependence on the ambient temperature, there was no good result. In this study, linear prediction methods and neural networks and fuzzy logic have been studied and emulated. An optimized fuzzy-wavelet prediction method is proposed to predict the price of electricity. In this method, in order to have a better prediction, the membership functions of the fuzzy regression along with the type of the wavelet transform filter have been optimized using the E.Coli Bacterial Foraging Optimization Algorithm. Then, to better compare this optimal method with other prediction methods including conventional linear prediction and neural network methods, they were analyzed with the same electricity price data. In fact, our fuzzy-wavelet method has a more desirable solution than previous methods. More precisely by choosing a suitable filter and a multiresolution processing method, the maximum error has improved by 13.6%, and the mean squared error has improved about 17.9%. In comparison with the fuzzy prediction method, our proposed method has a higher computational volume due to the use of wavelet transform as well as double use of fuzzy prediction. Due to the large number of layers and neurons used in it, the neural network method has a much higher computational volume than our fuzzy-wavelet method.

In this paper, using the Growing Neural Gas (GNG), the graph associated with a typed Persian character is extracted from the image. This graph learns the shape of the character and is used for character recognition. GNG, is one of the self-organized algorithms and a competitive learning approach. The main achievement of this method in addition to speed and accuracy is resistance to major challenges such as rotating, noise and deformation. This algorithm takes the image of a typed Persian character as input and returns the graph associated with that character.