Similarity or distance measures are core components used by distance-based clustering algorithms to cluster similar data points into the same clusters, while dissimilar distant placed different clusters. The performance of similarity is mostly addressed in two three-dimensional spaces, beyond which, best our knowledge, there no empirical study that has revealed behavior when dealing with high-dimensional datasets. To fill this gap, a technical framework proposed analyze, compare and...

Time series clustering is an important solution to various problems in numerous fields of research, including business, medical science, and finance. However, conventional algorithms are not practical for time data because they essentially designed static data. This impracticality results poor accuracy several systems. In this paper, a new hybrid algorithm proposed based on the similarity shape first grouped as subclusters time. The then merged using k -Medoids shape. model has two...

Digital image forgery is becoming easier to perform because of the rapid development various manipulation tools. Image splicing one most prevalent techniques. images had lost their trustability, and researches have exerted considerable effort regain such trustability by focusing mostly on algorithms. However, proposed algorithms are incapable handling high dimensionality redundancy in extracted features. Moreover, existing limited computational time. This study focuses improving detection...

Background: Tuberculosis (TB) is a major global health problem, which has been ranked as the second leading cause of death from an infectious disease worldwide.Diagnosis based on cultured specimens reference standard, however results take weeks to process.Scientists are looking for early detection strategies, remain cornerstone tuberculosis control.Consequently there need develop expert system that helps medical professionals accurately and quickly diagnose disease.Artificial Immune...

Time series clustering is a very effective approach in discovering valuable information various systems such as finance, embedded bio-sensor and genome. However, focusing on the efficiency scalability of these algorithms to deal with time seri

The recent extensive growth of data on the Web, has generated an enormous amount log records Web server databases. Applying usage mining techniques these vast amounts historical can discover potentially useful patterns and reveal user access behaviors site. Cluster analysis widely been applied to generate behavior models logs. Most off-line have problem decrease accuracy over time resulted new users joining or changes for existing in model-based approaches. This paper proposes a novel...

Dimensionality reduction (feature selection) is an important step in pattern recognition systems. Although there are different conventional approaches for feature selection, such as Principal Component Analysis, Random Projection, and Linear Discriminant selecting optimal, effective, robust features usually a difficult task. In this paper, new two-stage approach dimensionality proposed. This method based on one-dimensional two-dimensional spectrum diagrams of standard deviation minimum to...

Clustering is an unsupervised machine learning method that used both individually and as a part of the preprocessing stage for supervised methods. Due to its nature, clustering results have less accuracy compared learning. This article aims introduce new perspective in by defining approach data pruning. The also enables using multiple sets prototypes instead only one set improve accuracy. Consequently, this has potential be independently or prepare purified training step approach. An...

A major problem of pattern recognition systems is due to the large volume training datasets including duplicate and similar samples. In order overcome this problem, some dataset size reduction also dimensionality techniques have been introduced. The algorithms presently used for usually remove samples near centers classes or support vector between different classes. However, a class center include valuable information about characteristics important evaluating system efficiency. This paper...

Today, wide important advances in clustering time series have been obtained the field of data mining.A large part these successes are due to novel achieves dimensionality reduction and distance measurements data.However, addressing problem through conventional approach has not solved issue completely, especially when class label vague.In this paper, a two-level fuzzy strategy is employed order achieve objective.In first level, upon by symbolic representation, clustered high-level phase using...

With the extensive growth of data available on Internet, personalization this huge information becomes essential. Although, there are various techniques personalization, in paper we concentrate using mining algorithms to personalize web sites' usage data. This proposes an off-line model based that is generated by clustering algorithm.Then, will use users' transactions periodically change a dynamic-model.This proposed approach solve problem decrease accuracy models over time resulted new...

Nowadays, recommendation systems are definitely a necessity in the websites not just an auxiliary feature, especially for commercial and web sites with large information services. Recommendation use models constructed by applying statistical data mining approaches on derived from websites. In this paper we propose new hybrid approach that leverages usage domain of website to construct model. A model will be created clustering algorithm, then is adjusted based change behavior users or...

In the context of one-on-one instruction, reflective dialogues help students advance their learning and improve problem solving ability. The effectiveness instruction with respect to through dialogue is highlighted by researchers educators. However, little if any, known about how may lead improvement predict students’ This information can be extracted from large educational datasets using data mining techniques. Consequently, this study aims at USNA physics set applying a two-level...