Knowledge Discovery In Large Databases Research Articles

Time Series Analysis and Data Prediction of Large Databases: An Application to Electricity Demand Prediction

We evaluate statistical and machine learning methods for half-hourly 1-step-ahead electricity demand prediction using Australian electricity data. We show that the machine learning methods that use autocorrelation feature selection and BackPropagation Neural Networks, Linear Regression as prediction algorithms outperform the statistical methods Exponential Smoothing and also a number of baselines. We analyze the effect of day time on the prediction error and show that there are time-intervals associated with higher and lower errors and that the prediction methods also differ in their accuracy during the different time intervals. This analysis provides the foundation for construction a hybrid prediction model that achieved lower prediction error. We also show that an RBF neural network trained by genetic algorithm can achieved better prediction results than classic one. The aspect of increased transparency of networks through genetic evolution development features and granular computation is another essential topic promoted by knowledge discovery in large databases.

PHD: an efficient data clustering scheme using partition space technique for knowledge discovery in large databases

Rapid technological advances imply that the amount of data stored in databases is rising very fast. However, data mining can discover helpful implicit information in large databases. How to detect the implicit and useful information with lower time cost, high correctness, high noise filtering rate and fit for large databases is of priority concern in data mining, specifying why considerable clustering schemes have been proposed in recent decades. This investigation presents a new data clustering approach called PHD, which is an enhanced version of KIDBSCAN. PHD is a hybrid density-based algorithm, which partitions the data set by K-means, and then clusters the resulting partitions with IDBSCAN. Finally, the closest pairs of clusters are merged until the natural number of clusters of data set is reached. Experimental results reveal that the proposed algorithm can perform the entire clustering, and efficiently reduce the run-time cost. They also indicate that the proposed new clustering algorithm conducts better than several existing well-known schemes such as the K-means, DBSCAN, IDBSCAN and KIDBSCAN algorithms. Consequently, the proposed PHD algorithm is efficient and effective for data clustering in large databases.

Data Mining And Soft Computing

The term data mining refers to information elicitation. On the other hand, soft computing deals with information processing. If these two key properties can be combined in a constructive way, then this formation can effectively be used for knowledge discovery in large databases. Referring to this synergetic combination, the basic merits of data mining and soft computing paradigms are pointed out and novel data mining implementation coupled to a soft computing approach for knowledge discovery is presented. Knowledge modeling by machine learning together with the computer experiments is described and the effectiveness of the machine learning approach employed is demonstrated.

Applying Association Rule Discovery Algorithm to Multipoint Linkage Analysis.

Knowledge discovery in large databases (KDD) is being performed in several application domains, for example, the analysis of sales data, and is expected to be applied to other domains. We propose a KDD approach to multipoint linkage analysis, which is a way of ordering loci on a chromosome. Strict multipoint linkage analysis based on maximum likelihood estimation is a computationally tough problem. So far various kinds of approximate methods have been implemented. Our method based on the discovery of association between genetic recombinations is so different from others that it is useful to recheck the result of them. In this paper, we describe how to apply the framework of association rule discovery to linkage analysis, and also discuss that filtering input data and interpretation of discovered rules after data mining are practically important as well as data mining process itself.