The k-means, one of the most widely used clustering algorithm, is not only faster in computation but also produces comparatively better clusters. However, it has two major downsides, first it is sensitive to initializekvalue and secondly, especially for larger datasets, the number of iterations could be very large, making it computationally hard. In order to address these issues, we proposed a scalable and cost-effective algorithm, calledR-k-means,which provides an optimized solution for better clustering large scale high-dimensional datasets. The algorithm first selectsO(R)initial points then reselectO(l)better initial points, using distance probability from dataset. These points are then again clustered intokinitial points. An empirical study in a controlled environment was conducted using both simulated and real datasets. Experimental results showed that the proposed approach outperformed as compared to the previous approaches when the size of data increases with increasing number of dimensions.
In the process of knowledge discovery, the reliability of results depends upon the effectiveness of attributes selected for decision. The “curse of dimensionality” refers to the phenomenon in which the excessive number of dimensions affect the analysis. In order to eradicate the curse of dimensionality in text analysis, we are proposing an ontology-based semantic measure for intelligent selection/reduction of features. Among the various text mining techniques, ontology-based mining has a significant contribution to the field. Ontology-based semantic measures, which are mathematical models used to find the similarity between various concepts in the ontology, have made a significant contribution to feature engineering. The proposed measure is an amalgamation of semantic similarity, relatedness, and distance. The measure allows performing in-depth analysis of various semantic relationships between concepts of the English language. The performance of the measure was evaluated against benchmarked dimension reduction techniques such as PCA. The results show improvement by reducing the size of dimensions up to 35%. The results were further evaluated by training a classifier to validate that features are not creating any underfit/overfit model.
Actionable Knowledge Discovery approaches to extract the business and technical significant actions/patterns to support direct decision making. These actions suggest how to transform an object from an undesirable status to a desirable status by incurring less cost and high profit. This article aims to propose a work that generates actionable patterns efficiently. It reduces the search space and number of iterations for attribute value change during action generation. Performance of the proposed method is compared with Yang’s method and OF-CEAMA on the basis of four parameters i.e. the total number of rules required for action generation, run time of the methods, the total number of generated actions, total net profit and time and space complexity. Experiments have been carried out on four datasets retrieved from UCI Machine learning repository. Experimental results show that the proposed work takes less time than the other two methods to extract actions for all datasets. Also, the number of rules required to generate actions are less than the other two methods. Results also suggest that a decrease in execution time does not compromise the information and proposed work generates the same actions and net profit. Moreover, the proposed work tries to transfer an object from undesired status to the desired status.
30 June 2017 in Sukkur IBA Journal of Computing and Mathematical Sciences
Distributed Machine Learning (DML) has gained its importance more than ever in this era of Big Data. There are a lot of challenges to scale machine learning techniques on distributed platforms. When it comes to scalability, improving the processor technology for high level computation of data is at its limit, however increasing machine nodes and distributing data along with computation looks as a viable solution. Different frameworks and platforms are available to solve DML problems. These platforms provide automated random data distribution of datasets which miss the power of user defined intelligent data partitioning based on domain knowledge. We have conducted an empirical study which uses an EEG Data Set collected through P300 Speller component of an ERP (Event Related Potential) which is widely used in BCI problems; it helps in translating the intention of subject w h i l e performing any cognitive task. EEG data contains noise due to waves generated by other activities in the brain which contaminates true P300Speller. Use of Machine Learning techniques could help in detecting errors made by P300 Speller. We are solving this classification problem by partitioning data into different chunks and preparing distributed models using Elastic CV Classifier. To present a case of optimizing distributed machine learning, we propose an intelligent user defined data partitioning approach that could impact on the accuracy of distributed machine learners on average. Our results show better average AUC as compared to average AUC obtained after applying random data partitioning which gives no control to user over data partitioning. It improves the average accuracy of distributed learner due to the domain specific intelligent partitioning by the user. Our customized approach achieves 0.66 AUC on individual sessions and 0.75 AUC on mixed sessions, whereas random / uncontrolled data distribution records 0.63 AUC.
1 January 2014 in International Journal of Bioscience, Biochemistry and Bioinformatics