Improving accuracy and speed of optimum-path forest classifier using combination of disjoint training subsets

The Optimum-Path Forest (OPF) classifier is a recent and promising method for pattern recognition, with a fast training algorithm and good accuracy results. Therefore, the investigation of a combining method for this kind of classifier can be important for many applications. In this paper we report a fast method to combine OPF-based classifiers trained with disjoint training subsets. Given a fixed number of subsets, the algorithm chooses random samples, without replacement, from the original training set. Each subset accuracy is improved by a learning procedure. The final decision is given by majority vote. Experiments with simulated and real data sets showed that the proposed combining method is more efficient and effective than naive approach provided some conditions. It was also showed that OPF training step runs faster for a series of small subsets than for the whole training set. The combining scheme was also designed to support parallel or distributed processing, speeding up the procedure even more. © 2011 Springer-Verlag.

Land use image classification through optimum-path forest clustering

Land use classification has been paramount in the last years, since we can identify illegal land use and also to monitor deforesting areas. Although one can find several research works in the literature that address this problem, we propose here the land use recognition by means of Optimum-Path Forest Clustering (OPF), which has never been applied to this context up to date. Experiments among Optimum-Path Forest, Mean Shift and K-Means demonstrated the robustness of OPF for automatic land use classification of images obtained by CBERS-2B and Ikonos-2 satellites. © 2011 IEEE.

Spatial clustering applied to health area

The significant volume of work accidents in the cities causes an expressive loss to society. The development of Spatial Data Mining technologies presents a new perspective for the extraction of knowledge from the correlation between conventional and spatial attributes. One of the most important techniques of the Spatial Data Mining is the Spatial Clustering, which clusters similar spatial objects to find a distribution of patterns, taking into account the geographical position of the objects. Applying this technique to the health area, will provide information that can contribute towards the planning of more adequate strategies for the prevention of work accidents. The original contribution of this work is to present an application of tools developed for Spatial Clustering which supply a set of graphic resources that have helped to discover knowledge and support for management in the work accidents area. © 2011 IEEE.

Post-processing association rules with clustering and objective measures

The post-processing of association rules is a difficult task, since a large number of patterns can be obtained. Many approaches have been developed to overcome this problem, as objective measures and clustering, which are respectively used to: (i) highlight the potentially interesting knowledge in domain; (ii) structure the domain, organizing the rules in groups that contain, somehow, similar knowledge. However, objective measures don't reduce nor organize the collection of rules, making the understanding of the domain difficult. On the other hand, clustering doesn't reduce the exploration space nor direct the user to find interesting knowledge, making the search for relevant knowledge not so easy. This work proposes the PAR-COM (Post-processing Association Rules with Clustering and Objective Measures) methodology that, combining clustering and objective measures, reduces the association rule exploration space directing the user to what is potentially interesting. Thereby, PAR-COM minimizes the user's effort during the post-processing process.

The accuracy potential of Galileo E5/E1 pseudoranges for surveying and mapping

In the paper we discuss the potential of the new Galileo signals for pseudorange based surveying and mapping in open areas under optimal reception conditions (open sky scenarios) and suboptimal ones (multipath created by moderate to thick tree coverage). The paper reviews the main features of the Galileo E5 AltBOC and E1 CBOC signals; describes the simulation strategy, models and algorithms to generate realistic E5 and E1 pseudoranges with and without multipath sources; describes the ionosphere modeling strategy, models and algorithms and discusses and presents the expected positioning accuracy and precision results. According to the simulations performed, pseudoranges can be extracted from the Galileo E5 AltBOC signals with tracking errors (1-σ level) ranging from 0.02 m (open sky scenarios) to 0.08 m (tree covered scenarios) whereas for the Galileo E1 CBOC signals the tracking errors range between 0.25 m to 2.00 m respectively. With these tracking errors and with the explicit estimation of the ionosphere parameters, simulations indicate real-time open sky cm-level horizontal positioning precisions and dm-level vertical ones and dm-level accuracies for both the horizontal and vertical position components.

Characterization of the temperature, load and damage effects using piezoelectric transducer patches based on fuzzy clustering

Structural Health Monitoring (SHM) denotes a system with the ability to detect and interpret adverse changes in a structure. One of the critical challenges for practical implementation of SHM system is the ability to detect damage under changing environmental conditions. This paper aims to characterize the temperature, load and damage effects in the sensor measurements obtained with piezoelectric transducer (PZT) patches. Data sets are collected on thin aluminum specimens under different environmental conditions and artificially induced damage states. The fuzzy clustering algorithm is used to organize the sensor measurements into a set of clusters, which can attribute the variation in sensor data due to temperature, load or any induced damage.

Electrical consumers data clustering through optimum-path forest

Non-technical losses identification has been paramount in the last decade. Since we have datasets with hundreds of legal and illegal profiles, one may have a method to group data into subprofiles in order to minimize the search for consumers that cause great frauds. In this context, a electric power company may be interested in to go deeper a specific profile of illegal consumer. In this paper, we introduce the Optimum-Path Forest (OPF) clustering technique to this task, and we evaluate the behavior of a dataset provided by a brazilian electric power company with different values of an OPF parameter. © 2011 IEEE.

QK-Means: A clustering technique based on community detection and K-Means for deployment of cluster head nodes

Wireless Sensor Networks (WSN) are a special kind of ad-hoc networks that is usually deployed in a monitoring field in order to detect some physical phenomenon. Due to the low dependability of individual nodes, small radio coverage and large areas to be monitored, the organization of nodes in small clusters is generally used. Moreover, a large number of WSN nodes is usually deployed in the monitoring area to increase WSN dependability. Therefore, the best cluster head positioning is a desirable characteristic in a WSN. In this paper, we propose a hybrid clustering algorithm based on community detection in complex networks and traditional K-means clustering technique: the QK-Means algorithm. Simulation results show that QK-Means detect communities and sub-communities thus lost message rate is decreased and WSN coverage is increased. © 2012 IEEE.

Labeling methods for association rule clustering

Although association mining has been highlighted in the last years, the huge number of rules that are generated hamper its use. To overcome this problem, many post-processing approaches were suggested, such as clustering, which organizes the rules in groups that contain, somehow, similar knowledge. Nevertheless, clustering can aid the user only if good descriptors be associated with each group. This is a relevant issue, since the labels will provide to the user a view of the topics to be explored, helping to guide its search. This is interesting, for example, when the user doesn't have, a priori, an idea where to start. Thus, the analysis of different labeling methods for association rule clustering is important. Considering the exposed arguments, this paper analyzes some labeling methods through two measures that are proposed. One of them, Precision, measures how much the methods can find labels that represent as accurately as possible the rules contained in its group and Repetition Frequency determines how the labels are distributed along the clusters. As a result, it was possible to identify the methods and the domain organizations with the best performances that can be applied in clusters of association rules.

Boosting optimum-path forest clustering through harmony Search and its applications for intrusion detection in computer networks

In this paper we propose a nature-inspired approach that can boost the Optimum-Path Forest (OPF) clustering algorithm by optimizing its parameters in a discrete lattice. The experiments in two public datasets have shown that the proposed algorithm can achieve similar parameters' values compared to the exhaustive search. Although, the proposed technique is faster than the traditional one, being interesting for intrusion detection in large scale traffic networks. © 2012 IEEE.

Automatic visual dictionary generation through Optimum-Path Forest clustering

Image categorization by means of bag of visual words has received increasing attention by the image processing and vision communities in the last years. In these approaches, each image is represented by invariant points of interest which are mapped to a Hilbert Space representing a visual dictionary which aims at comprising the most discriminative features in a set of images. Notwithstanding, the main problem of such approaches is to find a compact and representative dictionary. Finding such representative dictionary automatically with no user intervention is an even more difficult task. In this paper, we propose a method to automatically find such dictionary by employing a recent developed graph-based clustering algorithm called Optimum-Path Forest, which does not make any assumption about the visual dictionary's size and is more efficient and effective than the state-of-the-art techniques used for dictionary generation. © 2012 IEEE.

Intrusion detection in computer networks using optimum-path forest clustering

Nowadays, organizations face the problem of keeping their information protected, available and trustworthy. In this context, machine learning techniques have also been extensively applied to this task. Since manual labeling is very expensive, several works attempt to handle intrusion detection with traditional clustering algorithms. In this paper, we introduce a new pattern recognition technique called Optimum-Path Forest (OPF) clustering to this task. Experiments on three public datasets have showed that OPF classifier may be a suitable tool to detect intrusions on computer networks, since it outperformed some state-of-the-art unsupervised techniques. © 2012 IEEE.

Tooth embedding medium influences the accuracy of electronic apex locator

This study evaluated the influence of tooth embedding media on the accuracy ofan electronic apex locator. The root canal length of 20 human mandibular canines was measured by inserting a 15 K-file into the root canal up to the apical foramen. The distance was measured with a digital caliper. The embedding media evaluated were alginate, saline, floral foam or gauze soaked in saline. Electronic root canal length measurement was performed with Root ZX II. Data were analysed using ANOVA for repeated measurements and Tukey test, at a significance level of 5%. There was no difference between the actual root canal length measurement and the electronic reading recorded with alginate medium. The readings obtained with the other media differed from the actual root canal length measurements. Alginate provided greater accuracy in electronic root canal length determination by Root ZX II than saline, floral foam and gauze.

Metrics to support the evaluation of association rule clustering

Many topics related to association mining have received attention in the research community, especially the ones focused on the discovery of interesting knowledge. A promising approach, related to this topic, is the application of clustering in the pre-processing step to aid the user to find the relevant associative patterns of the domain. In this paper, we propose nine metrics to support the evaluation of this kind of approach. The metrics are important since they provide criteria to: (a) analyze the methodologies, (b) identify their positive and negative aspects, (c) carry out comparisons among them and, therefore, (d) help the users to select the most suitable solution for their problems. Some experiments were done in order to present how the metrics can be used and their usefulness. © 2013 Springer-Verlag GmbH.

Comparison of the accuracy for three dental impression techniques and index: An in vitro study

Objectives: This in vitro study compared the dimensional accuracy of stone index (I) and three impression techniques: tapered impression copings (T), squared impression copings (S) and modified squared impression copings (MS) for implant-supported prostheses. Methods: A master cast, with four parallel implant abutment analogs and a passive framework, were fabricated. Vinyl polysiloxane impression material was used for all impressions with two metal stock trays (open and closed tray). Four groups (I, T, S and MS) were tested (n = 5). A metallic framework was seated on each of the casts, one abutment screw was tightened, and the gap between the analog of implant and the framework was measured with a stereomicroscope. The groups' measurements (80 gap values) were analyzed using software (LeicaQWin - Leica Imaging Systems Ltd.) that received the images of a video camera coupled to a Leica stereomicroscope at 100× magnification. The results were statistically analyzed with Kruskal-Wallis One Way ANOVA on Ranks test followed by Dunn's Method, 0.05. Results: The mean values of abutment/framework interface gaps were: Master Cast = 32 μm (SD 2); Group I = 45 μm (SD 3); Group T = 78 μm (SD 25); Group S = 134 μm (SD 30); Group MS = 143 μm (SD 27). No significant difference was detected among Index and Master Cast (P = .05). Conclusion: Under the limitations of this study, it could be suggested that a more accurate working cast is possible using tapered impression copings techniques and stone index. © 2013 Japan Prosthodontic Society.