Data Mining - Concepts and Techniques.

We explore the effect of dimensionality on the nearest neighbor problem. We show that under a broad set of conditions (much broader than independent and identically distributed dimensions), as dimensionality increases, the distance to the nearest data point approaches the distance to the farthest data point. To provide a practical perspective, we present empirical results on both real and synthetic data sets that demonstrate that this effect can occur for as few as 10-15 dimensions. These results should not be interpreted to mean that high-dimensional indexing is never meaningful; we illustrate this point by identifying some high-dimensional workloads for which this effect does not occur. However, our results do emphasize that the methodology used almost universally in the database literature to evaluate high-dimensional indexing techniques is flawed, and should be modified. In particular, most such techniques proposed in the literature are not evaluated versus simple linear scan, and are evaluated over workloads for which nearest neighbor is not meaningful. Often, even the reported experiments, when analyzed carefully, show that linear scan would outperform the techniques being proposed on the workloads studied in high (10-15) dimensionality!.

When is nearest neighbor meaningful

In this paper, we investigate several state-of-the-art machine-learning methods for automated classification of clustered microcalcifications (MCs). The classifier is part of a computer-aided diagnosis (CADx) scheme that is aimed to assisting radiologists in making more accurate diagnoses of breast cancer on mammograms. The methods we considered were: support vector machine (SVM), kernel Fisher discriminant (KFD), relevance vector machine (RVM), and committee machines (ensemble averaging and AdaBoost), of which most have been developed recently in statistical learning theory. We formulated differentiation of malignant from benign MCs as a supervised learning problem, and applied these learning methods to develop the classification algorithm. As input, these methods used image features automatically extracted from clustered MCs. We tested these methods using a database of 697 clinical mammograms from 386 cases, which included a wide spectrum of difficult-to-classify cases. We analyzed the distribution of the cases in this database using the multidimensional scaling technique, which reveals that in the feature space the malignant cases are not trivially separable from the benign ones. We used receiver operating characteristic (ROC) analysis to evaluate and to compare classification performance by the different methods. In addition, we also investigated how to combine information from multiple-view mammograms of the same case so that the best decision can be made by a classifier. In our experiments, the kernel-based methods (i.e., SVM, KFD, and RVM) yielded the best performance (A/sub z/=0.85, SVM), significantly outperforming a well-established, clinically-proven CADx approach that is based on neural network (A/sub z/=0.80).

A study on several Machine-learning methods for classification of Malignant and benign clustered microcalcifications

Associative classification mining is a promising approach in data mining that utilizes the association rule discovery techniques to construct classification systems, also known as associative classifiers. In the last few years, a number of associative classification algorithms have been proposed, i.e. CPAR, CMAR, MCAR, MMAC and others. These algorithms employ several different rule discovery, rule ranking, rule pruning, rule prediction and rule evaluation methods. This paper focuses on surveying and comparing the state-of-the-art associative classification techniques with regards to the above criteria. Finally, future directions in associative classification, such as incremental learning and mining low-quality data sets, are also highlighted in this paper.

/pdf/a-review-of-associative-classification-mining-19wba0gf74.pdf

A review of associative classification mining

Limited energy supply is one of the major constraints in wireless sensor networks. A feasible strategy is to aggressively reduce the spatial sampling rate of sensors, that is, the density of the measure points in a field. By properly scheduling, we want to retain the high fidelity of data collection. In this paper, we propose a data collection method that is based on a careful analysis of the surveillance data reported by the sensors. By exploring the spatial correlation of sensing data, we dynamically partition the sensor nodes into clusters so that the sensors in the same cluster have similar surveillance time series. They can share the workload of data collection in the future since their future readings may likely be similar. Furthermore, during a short-time period, a sensor may report similar readings. Such a correlation in the data reported from the same sensor is called temporal correlation, which can be explored to further save energy. We develop a generic framework to address several important technical challenges, including how to partition the sensors into clusters, how to dynamically maintain the clusters in response to environmental changes, how to schedule the sensors in a cluster, how to explore temporal correlation, and how to restore the data in the sink with high fidelity. We conduct an extensive empirical study to test our method using both a real test bed system and a large-scale synthetic data set.

/pdf/an-energy-efficient-data-collection-framework-for-wireless-4avf8teqyx.pdf

An Energy-Efficient Data Collection Framework for Wireless Sensor Networks by Exploiting Spatiotemporal Correlation

Breast cancer represents the second leading cause of cancer deaths in women today and it is the most common type of cancer in women. This paper presents some experiments for tumour detection in digital mammography. We investigate the use of different data mining techniques, neural networks and association rule mining, for anomaly detection and classification. The results show that the two approaches performed well, obtaining a classification accuracy reaching over 70% percent for both techniques. Moreover, the experiments we conducted demonstrate the use and effectiveness of association rule mining in image categorization.

https://webdocs.cs.ualberta.ca/~zaiane/postscript/mdmkdd01.pdf

Application of data mining techniques for medical image classification

This paper proposes a new classification method based on association rule mining. This association rule-based classifier is experimented on a real dataset; a database of medical images. The system we propose consists of: a preprocessing phase, a phase for mining the resulted transactional database, and a final phase to organize the resulted association rules in a classification model. The experimental results show that the method performs well reaching over 80% in accuracy. Moreover, this paper illustrates, by comparison to other published research, how important the data cleaning phase is in building an accurate data mining architecture for image classification.

/pdf/mammography-classification-by-an-association-rule-based-3q3354j8em.pdf

Mammography classification by an association rule-based classifier

This paper presents two classification systems for medical images based on association rule mining. The system we propose consists of: a pre-processing phase, a phase for mining the resulted transactional database, and a final phase to organize the resulted association rules in a classification model. The experimental results show that the method performs well, reaching over 80% in accuracy. Moreover, this paper illustrates how important the data cleaning phase is in building an accurate data mining architecture for image classification.

Associative classifiers for medical images

Wireless sensor networks consist of nodes with the ability to measure, store, and process data, as well as to communicate wirelessly with nodes located in their wireless range. Users can issue queries over the network, e.g., retrieve information from nodes within a specified region, in applications such as environmental monitoring. Since the sensors have typically only a limited power supply, energy-efficient processing of the queries over the network is an important issue. In this paper, we introduce a general framework for distributed processing of spatio-temporal queries in a sensor network that has two main phases: (1) routing the query to the spatial area specified in the query; (2) collecting and processing the information from the nodes relevant to the query. Within this framework, different algorithms can be designed independently for each of the two phases. We also propose novel algorithms for this framework, one for the first phase and two for the second phase. In an extensive experimental evaluation we study the performance of these algorithms in terms of energy consumption, under varying conditions. The results allow us to recommend the most energy efficient solution, given a network and a spatiotemporal query.

A framework for spatio-temporal query processing over wireless sensor networks

Sensor networks are an emerging technology that provide new means to monitor and understand various phenomena. Nodes in a sensor network typically have a limited power supply, thus energy-efficient processing of the queries over the network is a critical issue. We propose analytical models to evaluate the performance of three methods for processing historical spatio-temporal queries in sensor networks. The models allow us to better understand the tradeoffs of the investigated methods, as well as to recommend the most energy efficient one at query time.

/pdf/an-analysis-of-spatio-temporal-query-processing-in-sensor-dwfpx7kkx2.pdf

Alexandru Coman

Papers

Application of data mining techniques for medical image classification

Mammography classification by an association rule-based classifier

Associative classifiers for medical images

A framework for spatio-temporal query processing over wireless sensor networks

An Analysis of Spatio-Temporal Query Processing in Sensor Networks