Papers by Dost Muhammad Khan
Intelligent mobile agents are today accepted as powerful tools for data mining in a distributed e... more Intelligent mobile agents are today accepted as powerful tools for data mining in a distributed environment. The use of data mining algorithms further beefs up the intelligence in software agents. Knowledge discovery and data mining algorithms are applied to ...
Advances in mobile technology have enabled a large number of applications to be developed that ma... more Advances in mobile technology have enabled a large number of applications to be developed that may be employed by individual; this also produces the ‘usability’ as a major factor for development. In mobile technology, mobile phones are being used by common people and behavior of a user is highly influenced by his background and technical knowledge. On the other hand the focus of the mobile application development companies is on different features on mobile applications not on their ‘usability’. It is also difficult for developers to produce such an application that can equally satisfy all types of users; therefore, we highlight the issue of ‘usability’ in interactive mobile applications. In this research paper we pursue the usability issues that are encountered by common users of smartphone with no or less technical background and furthermore, we purpose a novel framework that can help the user to visualize applications according to his requirements.
ABSTRACT Cloud computing has developed from being a gifted commerce idea to one of the top geared... more ABSTRACT Cloud computing has developed from being a gifted commerce idea to one of the top geared sector of the Information Technology. Now, declined organizations are progressively introducing themselves in this technology in order to achieve reliable services at minimal cost. But as small and medium size business are looking forward to adopt least economical computing resources for their business applications, there is a need to identify all the issues while deploying it. The paper highlights some of most critical issues along with some mitigating steps in order to achieve rewarding deployment. This also describes some future development work of under laying concept.
Procedia Computer Science, 2013
The pattern extraction and discovery of useful information from a dataset are the foremost purpos... more The pattern extraction and discovery of useful information from a dataset are the foremost purposes of data mining; the multiple attempts and strong beliefs in the development and the formulation of the unified data mining frameworks that would answer to the fundamental versions related to the discovery of knowledge. In this paper we are presenting a novel unified framework for data mining conceptualized through the composite functions. The framework is further illustrated with a variety of real life datasets using different data mining algorithms.
Morphological filters erosion and dilation plays an important role for highlighting the buildings... more Morphological filters erosion and dilation plays an important role for highlighting the buildings and roads areas in aerial images. These filters give better results for classification of particular building and road areas in Images which are our region of interest. Image processing is a dynamic field and in recent years it is applied in many areas like medical sciences, remote sensing, military defense system, physical sciences and aerospace for useful purposes. This research paper focuses the study of highly resoluted aerial image of populated area. By applying different image processing techniques on aerial images like red band extraction, edge detection, enhanced gray level, texture features, red band replacement, binary threshold and finding the objects and classes of the image. By considering a particular range of buildings and roads objects then reconstruct the subjected images. At the end these buildings and road areas are highlighted by applying morphological filter erosion and dilation. We have taken the aerial image of Lahore Canal bank Road by Google earth and implemented all the above discussed techniques on it. The implemented techniques highlighted the building and road areas with red color.
Abstract Clustering is a technique in data mining to find interesting patterns in a given dataset... more Abstract Clustering is a technique in data mining to find interesting patterns in a given dataset. A large dataset is grouped into clusters of smaller sets of similar data using k-means algorithm. Initial centroids are required as input parameters when using k-means ...
Clustering is a technique in data mining to find interesting patterns in a given dataset. A large... more Clustering is a technique in data mining to find interesting patterns in a given dataset. A large dataset is grouped into clusters of smaller sets of similar data using k-means algorithm. Initial centroids are required as input parameters when using k-means clustering algorithm. There are different methods to choose initial centroids, from actual sample datapoints of a dataset. These methods are often implemented through intelligent agents, as the later are very commonly used in distributed networks given that they are not cumbersome for the network traffic. More over, they overcome network latency, operate in heterogeneous environment and possess fault-tolerant behavior. A multiagent system (MAS) is proposed in this research paper for the generation of initial centroids using actual sample datapoints. This multiagent system comprises four agents of k-means clustering algorithm using different methods namely Range, Random number, Outlier and Inlier for the generation of initial centroids.
International Journal of Information Processing and Management, 2010
Intelligent agents are today accepted as powerful tools for data mining in a distributed environm... more Intelligent agents are today accepted as powerful tools for data mining in a distributed environment. Artificial Intelligence (AI) algorithms are utilized to render software agents more intelligent. Knowledge discovery and data mining methods have been applied to discover hidden patterns and relations in complex datasets where clustering is one of the most important approaches used for this purpose. There exists a number of clustering algorithms among which the K-means algorithm is commonly used to find clusters due to its simplicity of implementation and fast execution. It appears extensively in the machine learning literature and in most data mining suite of tools. The algorithm is significantly sensitive to the selection of initial centroids. In this paper, we will present an agent oriented approach for implementation of the Range Method of initial centroids in K-means data mining algorithm. This Range Method is based on the actual sample datapoints. We have tested this method with both Euclidean and City Block (Manhattan) distances formulae on a different number of real life datasets.
Medical datasets hold huge number of records about the patients, the doctors and the diseases. Th... more Medical datasets hold huge number of records about the patients, the doctors and the diseases. The extraction of useful information which will provide knowledge in decision making process for the diagnosis and treatment of the diseases are becoming increasingly determinant. Knowledge Discovery and data mining make use of Artificial Intelligence (AI) algorithms which are applied to discover hidden patterns and relations in complex datasets using intelligent agents. The existing data mining algorithms and techniques are designed to solve the individual problems, such as classification or clustering. Up till now, no unifying theory is developed.Among the different algorithms in data mining for prediction, classification, interpretation and visualization, ‘k-means clustering’, ‘Decision Trees (C4.5)’, ‘Neural Network (NNs)’ and ‘Data Visualization (2D or 3D scattered graphs)’ algorithms are frequently utilized in data mining tools. The choice of the algorithm depends on the intended use of extracted knowledge.In this paper, the mentioned algorithms are unified into a tool, called Unified Medical Data Miner (UMDM) that will enable prediction, classification, interpretation and visualization on a diabetes dataset.
As a matter of fact there is unanimity that data mining is not a single-step process and the disc... more As a matter of fact there is unanimity that data mining is not a single-step process and the discover of knowledge from a dataset is the result of successive processes called multi-steps. The current data mining tools are designed to solve discrete consecutive tasks, such as classification or clustering, hence the tools turn out to be a single-step process and fail to produce the knowledge. Furthermore, in single-step tools the extraction of knowledge depends on the right choice of algorithms to apply and how to analyze the output, because most of them are generic and there is no context specific logic that is attached to the application. The choice of the algorithm remains ad-hoc in many data mining tools. The scientific community is very much conscious about this problematical issue and faced multiple challenges in establishing consensus over a unified data mining theory (UDMT) based on multi-step data mining processes. In this paper we draw a comparison between a single-step and multi-steps data mining tools. In singlestep data mining tools the selection of algorithms is on ad-hoc based which is inadequate to produce the 'knowledge'. On the other hand the multi-step data mining tool where the selection of the algorithms depends on the nature of the data provides the 'knowledge' to the user.
Uploads
Papers by Dost Muhammad Khan