Sharif Digital Repository

BCI is an interface between brain and machine, particularly computer which translates brain signals into understandable instructions for machine. BCI records signals and determines what the subject is doing or thinking. BCI in the point of view of pattern recognition is a classification problem. For this aim, different tasks are referred to different classes. The more number of classes, the higher complexity we encounter in classification so surveying of different kinds of features, feature selection and reduction methods have highly importance. In this project we want to design a 4-class classification that each class is referred to a direction of wrist movement. During the time that the... 

Data mining (DM) is one of the newest techniques which is used in decision making. DM helps the specialist to find the valuable knowledge which is hidden in data, with different techniques. DM has many research areas such as scientific research, medical fields, health care, fraud detection, marketing and customer relationship, sport, and games, and where ever there is data. Unfortunately, in our country, DM is not engaged seriously and there are fallacies of DM that weaken its efficiency. In this thesis, which is prepared in two sections, at the first section different techniques of DM are described and in the second section DM process is performed for a real world problem. For this purpose,... 

Coreference resolution is an essential step toward understanding discourses, and it is needed by many NLP tasks such as summarization, machine translation, question answering, etc. Pronoun resolution is a major and challenging subpart of coreference resolution, in which only the resolution of pronouns is considered. The existing coreference resolution approaches can be classified into two broad categories: linguistic and machine learning approaches. Linguistic approaches need a lot of linguistic information for the resolution process. Acquisition of such information is an error- prone and time-consuming process. In contrast, learning approaches need less linguistic information and provide... 

The classification of different tumor types is of great importance in cancer diagnosis and drug discovery. However, most previous cancer classification studies are clinical based and have limited diagnostic ability. Cancer classification using gene expression data is known to contain the keys for addressing the fundamental problems relating to cancer diagnosis. The recent advent of DNA microarray technique has made simultaneous monitoring of thousands of gene expressions possible. With this abundance of gene expression data, researchers have started to explore the possibilities of cancer classification using gene expression data and quite a number of Pattern Recognition approaches have been... 

Decision making often has different effects and results with unequal importance. Most of classifiers try to minimize the rate of misclassified instances. These classifiers assume equal costs for different misclassification types. However, this assumption is not true in many real world problems and different misclassification types have different costs. These differences can be applied by introducing the cost in the process of learning. In this manner, total cost of misclassification will be the evaluation metric of classification. In order to apply this metric to the problems, new learning algorithms are needed. Cost-sensitive learning is the related area of machine learning which deals with... 

DNA Microarray is the state-of-the-art technology in analyzing gene expression data. It has made it possible to measure expression levels of thousand of genes simultaneously. Microarray classification has been widely used in effective diagnosis of cancers and some other biological diseases. But the most challenging issue is the intense asymmetry between the dimensionality of genes and tissue samples which can wreck the classification performance. This dissertation will focus on gene selection and reduction solutions and presents a novel classification scheme which uses both gene selection and dimension reduction in its different stages. We have improved one of the recently proposed topology... 

In this project we tried to to solve the problem of clustring images of a large image database. Considering that there is no prior information regarding domain of the images, we will review unsupervised clustring methods. For this, we will discuss about image description vector and similarity measures. At last, our contribution will be about finding the optimum number of clusters with the least of user invervention. Results of runnig our method on a databse with 1000 images is reported and compared to a similar method named CLUE. Our result shows considerable improvements when user feedback taken to account.
 

Today many applications produce data stream that grow continuously over time. Classification is an important task in mining this data. One of the major challenges in classification of data streams is that the underlying concept of data may change over time; this needs updating the classification model. Many of traditional classification algorithms are adapted for classifying data streams. Evolutionary algorithms are stochastic methods that have been successfully applied to a wide range of optimization problems. Harmony search is a new evolutionary algorithm that convergence to global optimum and can be applied to discrete and continuous data. The evolutionary nature of classifying data... 

Content-based image retrieval (CBIR) is becoming an important field with the advance of multimedia and imaging technology everincreasingly. It makes use of image features, such as color, shape and texture, to index images with minimal human intervention. Among many retrieval features associated with CBIR, texture retrieval is one of the most powerful. Content-based image retrieval can also be utilized to locate medical images in large databases. In this research, we introduce a content-based approach to medical image retrieval. A case study, which describes the methodology of a CBIR system for retrieving digital human brain MRI database based on textural features retrieval, is then... 

The development of effective models for classification problems, such as the problem of selecting which credit applicants to accept, has been the subject of intense research for decades. Many static and dynamic methods, ranging from statistical classifiers to decision trees, nearest-neighbor methods, and neural networks, have already been proposed to tackle this problem and to assist decision making in the area of consumer and commercial credit. Given the profusion of modeling and data management techniques, it is often the case that which model has the more appropriate outputs in classification of the same problem. Among the stated methods although the latter, Neural Networks, is powerful... 

Semi-structured documents are a new type of textual documents which has gained a lot of attention nowadays to itself. A specific document modeling for boosting classifiers is needed more than ever which reflects major document specifications. The main goal of this thesis is presenting new adaptive model based on semi-structured documents features. We also aim to use some problem solving approaches such as Statistical approach, Machine Learning and few Algorithmic solutions. In some cases 10% precision optimization can be seen compare to the best approaches available nowadays  

In this thesis, the efficiency of bump modeling has been investigated on brain signals, in a variety of aspects including analysis, detection, classification and prediction. The aim of bump modeling is to provide an optimized representation of the signal in time-frequency domain. This would be done by discriminating oscillatory bursts from background signal and then showing them by half-ellipsoid functions called bump. Consequently, the problem of dealing with large numbers of parameters and hence complicated calculations, which are serious concerns in similar methods, can be overcome. This is in addition to the benefits of using time-frequency representation of the signal.The aim of bump... 

Traditional classification techniques of machine learning assume that data have stationary distributions. This assumption for recent challenges where tremendous amount of data are generated at unprecedented rates with evolving patterns, is not true anymore. Classification of data streams has become an important area of machine learning, as the number of applications facing these challenges increases. Examples of such data streams applications include text streams, surveillance video streams, credit card fraud detection, market basket analysis, information filtering, computer security, etc. An appropriate method for such problems should adapt to drifting concepts by revising and refining the... 

One of the most applicable fields in data mining is customer relationship management (CRM). CRM process includes four aspects: Customer identification, Customer attraction, Customer retention, and Customer development. Data mining can be a supportive tool for decision making in each of these CRM aspects. Huge volume of data and information corresponding to CRM that exists in companies' databases, has made sufficient potential for data mining process and discovering hidden knowledge. Importance of concepts like customer needs identification, customer retention, and increasing customer value for companies has made the need to use of data mining techniques more valuable. Saipa Yadak Co., as a... 

Computer-Aided Diagnosis (CAD) systems are widely used for detection of various kinds of abnormalities in mammography images. In this work, mass classification is investigated and its steps are explained in detail, for each step a main method is presented and other methods are also discussed. For mass segmentation a relatively new method based on level set and Morphological Component Analysis (MCA) is used.After this step, various kinds of features such as shape, geometrical, and textural ones are introduced. Moreover, a set of proposed features based on wavelet transformation,for this application are presented. The proposed features can describe margin and texture characterizations of a... 

Cardiovascular diseases are very common nowadays and are one of the main reasons of death. Being among the major types of these diseases, correct and in time diagnosis of Coronary Artery Disease (CAD) is very important. The best and most accurate CAD diagnosis method by now is recognized as Angiography, which has many side effects and is costly. Thus researchers are seeking for inexpensive, though still accurate, methods. Existing studies have used several features in collecting data from patients, while applying different data mining algorithms to increase accuracy. In this thesis, a data set is introduced which utilizes several new and effective features for CAD diagnosis, as well as a... 

Measuring emotions of music is one of the methods to determine music content. Music emotion detection is applicable in music retrieval, recognition of music genre and also music data management softwares. Music emotion is considered in different sciences such as physiology, psychology, musicology and engineering. First, we collected a database of different types of music with various emotions. These data have been labeled according to their emotions. In this project, four emotions (Angry, happy, relax and sad) have been used as labels based on Thayer’s two dimension emotion model. There are two basic steps for music emotion recognition similar to other recognition systems: Feature extraction... 

Cardio Vascular Diseases (CVD) is currently the most important cause of mortality in the world. Coronary artery disease (CAD-with 3 major epicardial vessels) is one of the significant CVD. Coronary angiography remains the gold standard for identifying the presence and absence of stenosis due to CAD. It is performed by inserting a catheter into the coronary artery, directly injecting radiopaque contrast material there and recording radiographic images in different view of arteries. Images of this invasive way show severity and area of stenosis.
A sign of CAD is endothelial dysfunction (inner artery layer) in peripheral arteries like the brachial artery. Due to a significant correlation... 

Event extraction is one of the important tasks in Natural Language Processing (NLP). Many NLP applications like question answering, information extraction and summarization need to have some knowledge about events of input documents. There are several definitions for events in NLP domains. In this dissertation, the event is viewed as an element in a network of temporal information. Therefore, the project has been based on the ISO-TimeML specification language, which is the standard scheme for temporal information processing in natural texts. Event extraction based on ISO-TimeML has been performed for a number of languages including English, French, Spanish, and Korean. However, for Persian... 

The main focus of this research is to resolve the semantic ambiguity in Persian. In this study, a semi-supervised machine learning method is proposed to choose the most proper meaning of a target word in the context. Several statistical methods are compared, and the most accurate one is chosen for developing a sense tagger. An initial seed data is built by searching collocation lists for each sense. After developing the sense tagger and initial seed set, a bootstrapping method is used to sense tag all occurences of a target word in corpus with 90% accuracy