Du, J. | Korkmaz, E.E. | Alhajj, R. | Barker, K.
Conference Object | 2005 | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)3587 LNAI , pp.346 - 355
In this paper, we present a linked-list based encoding scheme for multiple objectives based genetic algorithm (GA) to identify clusters in a partition. Our approach obtains the optimal partitions for all the possible numbers of clusters in the Pareto Optimal set returned by a single genetic GA run. The performance of the proposed approach has been tested using two well-known data sets, namely Iris and Ruspini. The obtained results are promising and demonstrate the applicability and effectiveness of the proposed approach. © Springer-Verlag Berlin Heidelberg 2005.
Gopalan, J. | Korkmaz, E. | Alhajj, R. | Barker, K.
Conference Object | 2005 | Proceedings - ICMLA 2005: Fourth International Conference on Machine Learning and Applications2005 , pp.331 - 336
Dividing a data set into a training set and a test set is a fundamental component in the pre-processing phase of data mining (DM). Effectively, the choice of the training set is an important factor in deriving good classification rules. Traditional approach for association rules mining divides the dataset into training set and test set based on statistical methods. In this paper, we highlight the weaknesses of the existing approach and hence propose a new methodology that employs genetic algorithm (GA) in the process. In our approach, the original dataset is divided into sample and validation sets. Then, GA is used to find an approp . . .riate split of the sample set into training and test sets. We demonstrate through experiments that using the obtained training set as the input to an association rules mining algorithm generates high accuracy classification rules. The rules are tested on the validation set for accuracy. The results are very satisfactory; they demonstrate the applicability and effectiveness of our approach. © 2005 IEEE
Altintas, B. | Serif, T.
Conference Object | 2011 | 17th European Wireless Conference 2011, EW 2011 , pp.681 - 685
Advances in mobile technologies and devices has changed the way users interact with devices and other users. These new interaction methods and services are offered by the help of intelligent sensing capabilities, using context, location and motion sensors. However, indoor location sensing is mostly achieved by utilizing radio signal (Wi-Fi, Bluetooth, GSM etc.) and nearest neighbor identification. The most common algorithm adopted for Received Signal Strength (RSS)-based location sensing is K Nearest Neighbor (KNN), which calculates K nearest neighboring points to mobile users (MUs). Accordingly, in this paper, we aim to improve the . . . KNN algorithm by enhancing the neighboring point selection by applying k-means clustering approach. In the proposed method, k-means clustering algorithm groups nearest neighbors according to their distance to mobile user. Then the closest group to the mobile user is used to calculate the MU's location. The evaluation results indicate that the performance of clustered KNN is closely tied to the number of clusters, number of neighbors to be clustered and the initiation of the center points in k-mean algorithm. © VDE VERLAG GMBH