Bagging k-dependence probabilistic networks: An alternative powerful fraud detection tool

Fraud is a global problem that has required more attention due to an accentuated expansion of modern technology and communication. When statistical techniques are used to detect fraud, whether a fraud detection model is accurate enough in order to provide correct classification of the case as a fraudulent or legitimate is a critical factor. In this context, the concept of bootstrap aggregating (bagging) arises. The basic idea is to generate multiple classifiers by obtaining the predicted values from the adjusted models to several replicated datasets and then combining them into a single predictive classification in order to improve the classification accuracy. In this paper, for the first time, we aim to present a pioneer study of the performance of the discrete and continuous k-dependence probabilistic networks within the context of bagging predictors classification. Via a large simulation study and various real datasets, we discovered that the probabilistic networks are a strong modeling option with high predictive capacity and with a high increment using the bagging procedure when compared to traditional techniques. (C) 2012 Elsevier Ltd. All rights reserved.

New Insights on Nontechnical Losses Characterization Through Evolutionary-Based Feature Selection

Although nontechnical losses automatic identification has been massively studied, the problem of selecting the most representative features in order to boost the identification accuracy and to characterize possible illegal consumers has not attracted much attention in this context. In this paper, we focus on this problem by reviewing three evolutionary-based techniques for feature selection, and we also introduce one of them in this context. The results demonstrated that selecting the most representative features can improve a lot of the classification accuracy of possible frauds in datasets composed by industrial and commercial profiles.

On the impact of disproportional samples in credit scoring models: An application to a Brazilian bank data

Statistical methods have been widely employed to assess the capabilities of credit scoring classification models in order to reduce the risk of wrong decisions when granting credit facilities to clients. The predictive quality of a classification model can be evaluated based on measures such as sensitivity, specificity, predictive values, accuracy, correlation coefficients and information theoretical measures, such as relative entropy and mutual information. In this paper we analyze the performance of a naive logistic regression model (Hosmer & Lemeshow, 1989) and a logistic regression with state-dependent sample selection model (Cramer, 2004) applied to simulated data. Also, as a case study, the methodology is illustrated on a data set extracted from a Brazilian bank portfolio. Our simulation results so far revealed that there is no statistically significant difference in terms of predictive capacity between the naive logistic regression models and the logistic regression with state-dependent sample selection models. However, there is strong difference between the distributions of the estimated default probabilities from these two statistical modeling techniques, with the naive logistic regression models always underestimating such probabilities, particularly in the presence of balanced samples. (C) 2012 Elsevier Ltd. All rights reserved.

Urban Expansion and its impact on urban agriculture - remote sensing based change analysis of Kizinga and Mzinga Valley - Dar Es Salaam, Tanzania

Urban agriculture is a phenomenon that can be observed world-wide, particularly in cities of devel- oping countries. It is contributing significantly to food security and food safety and has sustained livelihood of the urban and peri-urban low income dwe llers in developing countries for many years. Population increase due to rural-urban migration and natural - formal as well as informal - urbani- sation are competing with urban farming for available space and scarce water resources. A mul- titemporal and multisensoral urban change analysis over the period of 25 years (1982-2007) was performed in order to measure and visualise the urban expansion along the Kizinga and Mzinga valley in the south of Dar Es Salaam. Airphotos and VHR satellite data were analysed by using a combination of a composition of anisotropic textural measures and spectral information. The study revealed that unplanned built-up area is expanding continuously, and vegetation covers and agricultural lands decline at a fast rate. The validation showed that the overall classification accuracy varied depending on the database. The extracted built-up areas were used for visual in- terpretation mapping purposes and served as information source for another research project. The maps visualise an urban congestion and expansion of nearly 18% of the total analysed area that had taken place in the Kizinga valley between 1982 and 2007. The same development can be ob- served in the less developed and more remote Mzinga valley between 1981 and 2002. Both areas underwent fast changes where land prices still tend to go up and an influx of people both from rural and urban areas continuously increase the density with the consequence of increasing multiple land use interests.

Estimation of scots pine defoliation by the common pine sawfly (Diprion pini L.) using multi-temporal radar data

In 1998-2001 Finland suffered the most severe insect outbreak ever recorded, over 500,000 hectares. The outbreak was caused by the common pine sawfly (Diprion pini L.). The outbreak has continued in the study area, Palokangas, ever since. To find a good method to monitor this type of outbreaks, the purpose of this study was to examine the efficacy of multi-temporal ERS-2 and ENVISAT SAR imagery for estimating Scots pine (Pinus sylvestris L.) defoliation. Three methods were tested: unsupervised k-means clustering, supervised linear discriminant analysis (LDA) and logistic regression. In addition, I assessed if harvested areas could be differentiated from the defoliated forest using the same methods. Two different speckle filters were used to determine the effect of filtering on the SAR imagery and subsequent results. The logistic regression performed best, producing a classification accuracy of 81.6% (kappa 0.62) with two classes (no defoliation, >20% defoliation). LDA accuracy was with two classes at best 77.7% (kappa 0.54) and k-means 72.8 (0.46). In general, the largest speckle filter, 5 x 5 image window, performed best. When additional classes were added the accuracy was usually degraded on a step-by-step basis. The results were good, but because of the restrictions in the study they should be confirmed with independent data, before full conclusions can be made that results are reliable. The restrictions include the small size field data and, thus, the problems with accuracy assessment (no separate testing data) as well as the lack of meteorological data from the imaging dates.

A Food Recognition System for Diabetic Patients based on an Optimized Bag of Features Model

Computer vision-based food recognition could be used to estimate a meal's carbohydrate content for diabetic patients. This study proposes a methodology for automatic food recognition, based on the Bag of Features (BoF) model. An extensive technical investigation was conducted for the identification and optimization of the best performing components involved in the BoF architecture, as well as the estimation of the corresponding parameters. For the design and evaluation of the prototype system, a visual dataset with nearly 5,000 food images was created and organized into 11 classes. The optimized system computes dense local features, using the scale-invariant feature transform on the HSV color space, builds a visual dictionary of 10,000 visual words by using the hierarchical k-means clustering and finally classifies the food images with a linear support vector machine classifier. The system achieved classification accuracy of the order of 78%, thus proving the feasibility of the proposed approach in a very challenging image dataset.

Daytime Low Stratiform Cloud Detection on AVHRR Imagery

The near-real time retrieval of low stratiform cloud (LSC) coverage is of vital interest for such disciplines as meteorology, transport safety, economy and air quality. Within this scope, a novel methodology is proposed which provides the LSC occurrence probability estimates for a satellite scene. The algorithm is suited for the 1 × 1 km Advanced Very High Resolution Radiometer (AVHRR) data and was trained and validated against collocated SYNOP observations. Utilisation of these two combined data sources requires a formulation of constraints in order to discriminate cases where the LSC is overlaid by higher clouds. The LSC classification process is based on six features which are first converted to the integer form by step functions and combined by means of bitwise operations. Consequently, a set of values reflecting a unique combination of those features is derived which is further employed to extract the LSC occurrence probability estimates from the precomputed look-up vectors (LUV). Although the validation analyses confirmed good performance of the algorithm, some inevitable misclassification with other optically thick clouds were reported. Moreover, the comparison against Polar Platform System (PPS) cloud-type product revealed superior classification accuracy. From the temporal perspective, the acquired results reported a presence of diurnal and annual LSC probability cycles over Europe.

Urban expansion and its impact on urban agriculture – remote sensing based change analysis of Kizinga and Mzinga Valley – Dar Es Salaam, Tanzania

Urban agriculture is a phenomenon that can be observed world-wide, particularly in cities of devel-oping countries. It is contributing significantly to food security and food safety and has sustained livelihood of the urban and peri-urban low income dwellers in developing countries for many years. Population increase due to rural-urban migration and natural, coupled with formal as well as infor-mal urbanization are competing with urban farming for available space and scarce water resources. A multitemporal multisensoral urban change analysis over the period of 25 years (1982-2007) was performed in order to measure and visualize the urban expansion along the Kizinga and Mzinga valley in the South of Dar es Salaam. Airphotos and VHR satellite data were analyzed by using a combination of a composition of anisotropic textural measures and spectral information. The study revealed that unplanned built-up area is expanding continuously and vegetation covers and agricultural lands decline at a fast rate. The validation showed that the overall classification accuracy varied depending on the database. The extracted built-up areas were used for visual in-terpretation mapping purposes and served as information source for another research project. The maps visualize an urban congestion and expansion of nearly 18% of the total analyzed area that had taken place in the Kizinga valley between 1982 and 2007. The same development can be ob-served in the less developed and more remote Mzinga valley between 1981 and 2002. Both areas underwent fast changes where land prices still tend to go up and an influx of people both from rural and urban areas continuously increase density with the consequence of increasing multiple land use interests.

Automatic identification of discourse markers in dialogues: An in-depth study of like and well

The lexical items like and well can serve as discourse markers (DMs), but can also play numerous other roles, such as verb or adverb. Identifying the occurrences that function as DMs is an important step for language understanding by computers. In this study, automatic classifiers using lexical, prosodic/positional and sociolinguistic features are trained over transcribed dialogues, manually annotated with DM information. The resulting classifiers improve state-of-the-art performance of DM identification, at about 90% recall and 79% precision for like (84.5% accuracy, κ = 0.69), and 99% recall and 98% precision for well (97.5% accuracy, κ = 0.88). Automatic feature analysis shows that lexical collocations are the most reliable indicators, followed by prosodic/positional features, while sociolinguistic features are marginally useful for the identification of DM like and not useful for well. The differentiated processing of each type of DM improves classification accuracy, suggesting that these types should be treated individually.

The information bottleneck method for genome-wide association studies

In population studies, most current methods focus on identifying one outcome-related SNP at a time by testing for differences of genotype frequencies between disease and healthy groups or among different population groups. However, testing a great number of SNPs simultaneously has a problem of multiple testing and will give false-positive results. Although, this problem can be effectively dealt with through several approaches such as Bonferroni correction, permutation testing and false discovery rates, patterns of the joint effects by several genes, each with weak effect, might not be able to be determined. With the availability of high-throughput genotyping technology, searching for multiple scattered SNPs over the whole genome and modeling their joint effect on the target variable has become possible. Exhaustive search of all SNP subsets is computationally infeasible for millions of SNPs in a genome-wide study. Several effective feature selection methods combined with classification functions have been proposed to search for an optimal SNP subset among big data sets where the number of feature SNPs far exceeds the number of observations. ^ In this study, we take two steps to achieve the goal. First we selected 1000 SNPs through an effective filter method and then we performed a feature selection wrapped around a classifier to identify an optimal SNP subset for predicting disease. And also we developed a novel classification method-sequential information bottleneck method wrapped inside different search algorithms to identify an optimal subset of SNPs for classifying the outcome variable. This new method was compared with the classical linear discriminant analysis in terms of classification performance. Finally, we performed chi-square test to look at the relationship between each SNP and disease from another point of view. ^ In general, our results show that filtering features using harmononic mean of sensitivity and specificity(HMSS) through linear discriminant analysis (LDA) is better than using LDA training accuracy or mutual information in our study. Our results also demonstrate that exhaustive search of a small subset with one SNP, two SNPs or 3 SNP subset based on best 100 composite 2-SNPs can find an optimal subset and further inclusion of more SNPs through heuristic algorithm doesn't always increase the performance of SNP subsets. Although sequential forward floating selection can be applied to prevent from the nesting effect of forward selection, it does not always out-perform the latter due to overfitting from observing more complex subset states. ^ Our results also indicate that HMSS as a criterion to evaluate the classification ability of a function can be used in imbalanced data without modifying the original dataset as against classification accuracy. Our four studies suggest that Sequential Information Bottleneck(sIB), a new unsupervised technique, can be adopted to predict the outcome and its ability to detect the target status is superior to the traditional LDA in the study. ^ From our results we can see that the best test probability-HMSS for predicting CVD, stroke,CAD and psoriasis through sIB is 0.59406, 0.641815, 0.645315 and 0.678658, respectively. In terms of group prediction accuracy, the highest test accuracy of sIB for diagnosing a normal status among controls can reach 0.708999, 0.863216, 0.639918 and 0.850275 respectively in the four studies if the test accuracy among cases is required to be not less than 0.4. On the other hand, the highest test accuracy of sIB for diagnosing a disease among cases can reach 0.748644, 0.789916, 0.705701 and 0.749436 respectively in the four studies if the test accuracy among controls is required to be at least 0.4. ^ A further genome-wide association study through Chi square test shows that there are no significant SNPs detected at the cut-off level 9.09451E-08 in the Framingham heart study of CVD. Study results in WTCCC can only detect two significant SNPs that are associated with CAD. In the genome-wide study of psoriasis most of top 20 SNP markers with impressive classification accuracy are also significantly associated with the disease through chi-square test at the cut-off value 1.11E-07. ^ Although our classification methods can achieve high accuracy in the study, complete descriptions of those classification results(95% confidence interval or statistical test of differences) require more cost-effective methods or efficient computing system, both of which can't be accomplished currently in our genome-wide study. We should also note that the purpose of this study is to identify subsets of SNPs with high prediction ability and those SNPs with good discriminant power are not necessary to be causal markers for the disease.^

TIME SERIES ANALYSIS AS INPUT FOR PREDICTIVE MODELING: PREDICTING CARDIAC ARREST IN A PEDIATRIC INTENSIVE CARE UNIT

The first manuscript, entitled "Time-Series Analysis as Input for Clinical Predictive Modeling: Modeling Cardiac Arrest in a Pediatric ICU" lays out the theoretical background for the project. There are several core concepts presented in this paper. First, traditional multivariate models (where each variable is represented by only one value) provide single point-in-time snapshots of patient status: they are incapable of characterizing deterioration. Since deterioration is consistently identified as a precursor to cardiac arrests, we maintain that the traditional multivariate paradigm is insufficient for predicting arrests. We identify time series analysis as a method capable of characterizing deterioration in an objective, mathematical fashion, and describe how to build a general foundation for predictive modeling using time series analysis results as latent variables. Building a solid foundation for any given modeling task involves addressing a number of issues during the design phase. These include selecting the proper candidate features on which to base the model, and selecting the most appropriate tool to measure them. We also identified several unique design issues that are introduced when time series data elements are added to the set of candidate features. One such issue is in defining the duration and resolution of time series elements required to sufficiently characterize the time series phenomena being considered as candidate features for the predictive model. Once the duration and resolution are established, there must also be explicit mathematical or statistical operations that produce the time series analysis result to be used as a latent candidate feature. In synthesizing the comprehensive framework for building a predictive model based on time series data elements, we identified at least four classes of data that can be used in the model design. The first two classes are shared with traditional multivariate models: multivariate data and clinical latent features. Multivariate data is represented by the standard one value per variable paradigm and is widely employed in a host of clinical models and tools. These are often represented by a number present in a given cell of a table. Clinical latent features derived, rather than directly measured, data elements that more accurately represent a particular clinical phenomenon than any of the directly measured data elements in isolation. The second two classes are unique to the time series data elements. The first of these is the raw data elements. These are represented by multiple values per variable, and constitute the measured observations that are typically available to end users when they review time series data. These are often represented as dots on a graph. The final class of data results from performing time series analysis. This class of data represents the fundamental concept on which our hypothesis is based. The specific statistical or mathematical operations are up to the modeler to determine, but we generally recommend that a variety of analyses be performed in order to maximize the likelihood that a representation of the time series data elements is produced that is able to distinguish between two or more classes of outcomes. The second manuscript, entitled "Building Clinical Prediction Models Using Time Series Data: Modeling Cardiac Arrest in a Pediatric ICU" provides a detailed description, start to finish, of the methods required to prepare the data, build, and validate a predictive model that uses the time series data elements determined in the first paper. One of the fundamental tenets of the second paper is that manual implementations of time series based models are unfeasible due to the relatively large number of data elements and the complexity of preprocessing that must occur before data can be presented to the model. Each of the seventeen steps is analyzed from the perspective of how it may be automated, when necessary. We identify the general objectives and available strategies of each of the steps, and we present our rationale for choosing a specific strategy for each step in the case of predicting cardiac arrest in a pediatric intensive care unit. Another issue brought to light by the second paper is that the individual steps required to use time series data for predictive modeling are more numerous and more complex than those used for modeling with traditional multivariate data. Even after complexities attributable to the design phase (addressed in our first paper) have been accounted for, the management and manipulation of the time series elements (the preprocessing steps in particular) are issues that are not present in a traditional multivariate modeling paradigm. In our methods, we present the issues that arise from the time series data elements: defining a reference time; imputing and reducing time series data in order to conform to a predefined structure that was specified during the design phase; and normalizing variable families rather than individual variable instances. The final manuscript, entitled: "Using Time-Series Analysis to Predict Cardiac Arrest in a Pediatric Intensive Care Unit" presents the results that were obtained by applying the theoretical construct and its associated methods (detailed in the first two papers) to the case of cardiac arrest prediction in a pediatric intensive care unit. Our results showed that utilizing the trend analysis from the time series data elements reduced the number of classification errors by 73%. The area under the Receiver Operating Characteristic curve increased from a baseline of 87% to 98% by including the trend analysis. In addition to the performance measures, we were also able to demonstrate that adding raw time series data elements without their associated trend analyses improved classification accuracy as compared to the baseline multivariate model, but diminished classification accuracy as compared to when just the trend analysis features were added (ie, without adding the raw time series data elements). We believe this phenomenon was largely attributable to overfitting, which is known to increase as the ratio of candidate features to class examples rises. Furthermore, although we employed several feature reduction strategies to counteract the overfitting problem, they failed to improve the performance beyond that which was achieved by exclusion of the raw time series elements. Finally, our data demonstrated that pulse oximetry and systolic blood pressure readings tend to start diminishing about 10-20 minutes before an arrest, whereas heart rates tend to diminish rapidly less than 5 minutes before an arrest.

Victorian Dominant Land Cover, 2009-2013

There is a demand for regularly updated, broad-scale, accurate land cover information in Victoria from multiple stakeholders. This paper documents the methods used to generate an annual dominant land cover (DLC) map for Victoria, Australia from 2009 to 2013. Vegetation phenology parameters derived from an annual time series of the Moderate Resolution Imaging Spectroradiometer Vegetation Indices 16-day 250 m (MOD13Q1) product were used to generate annual DLC maps, using a three-tiered hierarchical classification scheme. Classification accuracy at the broadest (primary) class level was over 91% for all years, while it ranged from 72 to 81% at the secondary class level. The most detailed class level (tertiary) had accuracy levels ranging from 61 to 68%. The approach used was able to accommodate variable climatic conditions, which had substantial impacts on vegetation growth patterns and agricultural production across the state between both regions and years. The production of an annual dataset with complete spatial coverage for Victoria provides a reliable base data set with an accuracy that is fit-for-purpose for many applications.

Regularized logistic regression and multi-objective variable selection for classifying MEG data

This paper addresses the question of maximizing classifier accuracy for classifying task-related mental activity from Magnetoencelophalography (MEG) data. We propose the use of different sources of information and introduce an automatic channel selection procedure. To determine an informative set of channels, our approach combines a variety of machine learning algorithms: feature subset selection methods, classifiers based on regularized logistic regression, information fusion, and multiobjective optimization based on probabilistic modeling of the search space. The experimental results show that our proposal is able to improve classification accuracy compared to approaches whose classifiers use only one type of MEG information or for which the set of channels is fixed a priori.

Collaborative data stream mining in ubiquitous environments using dynamic classifier selection

In ubiquitous data stream mining applications, different devices often aim to learn concepts that are similar to some extent. In these applications, such as spam filtering or news recommendation, the data stream underlying concept (e.g., interesting mail/news) is likely to change over time. Therefore, the resultant model must be continuously adapted to such changes. This paper presents a novel Collaborative Data Stream Mining (Coll-Stream) approach that explores the similarities in the knowledge available from other devices to improve local classification accuracy. Coll-Stream integrates the community knowledge using an ensemble method where the classifiers are selected and weighted based on their local accuracy for different partitions of the feature space. We evaluate Coll-Stream classification accuracy in situations with concept drift, noise, partition granularity and concept similarity in relation to the local underlying concept. The experimental results show that Coll-Stream resultant model achieves stability and accuracy in a variety of situations using both synthetic and real world datasets.

Data mining for the diagnosis of type 2 diabetes

Diabetes is the most common disease nowadays in all populations and in all age groups. diabetes contributing to heart disease, increases the risks of developing kidney disease, blindness, nerve damage, and blood vessel damage. Diabetes disease diagnosis via proper interpretation of the diabetes data is an important classification problem. Different techniques of artificial intelligence has been applied to diabetes problem. The purpose of this study is apply the artificial metaplasticity on multilayer perceptron (AMMLP) as a data mining (DM) technique for the diabetes disease diagnosis. The Pima Indians diabetes was used to test the proposed model AMMLP. The results obtained by AMMLP were compared with decision tree (DT), Bayesian classifier (BC) and other algorithms, recently proposed by other researchers, that were applied to the same database. The robustness of the algorithms are examined using classification accuracy, analysis of sensitivity and specificity, confusion matrix. The results obtained by AMMLP are superior to obtained by DT and BC.