Classification of the cardiotocogram data for anticipation of fetal risks using machine learning techniques

The aim of the research is evaluating the classification performances of eight different machine-learning methods on the antepartum cardiotocography (CTG) data. The classification is necessary to predict newborn health, especially for the critical cases. Cardiotocography is used for assisting the obstetricians’ to obtain detailed information during the pregnancy as a technique of measuring fetal well-being, essentially in pregnant women having potential complications. The obstetricians describe CTG shortly as a continuous electronic record of the baby's heart rate took from the mother's abdomen. The acquired information is necessary to visualize unhealthiness of the embryo and gives an opportunity for early intervention prior to happening a permanent impairment to the embryo. The aim of the machine learning methods is by using attributes of data obtained from the uterine contraction (UC) and fetal heart rate (FHR) signals to classify as pathological or normal. The dataset contains 1831 instances with 21 attributes, examined by applying the methods. In the paper, the highest accuracy displayed as 99.2%.     

Prediction of risk factors of cyberbullying-related words in Korea: Application of data mining using social big data

The study examined a decision tree analysis using social big data to conduct the prediction model on types of risk factors related to cyberbullying in Korea. The study conducted an analysis of 103,212 buzzes that had noted causes of cyberbullying and data were collected from 227 online channels, such as news websites, blogs, online groups, social network services, and online bulletin boards. Using opinion-mining method and decision tree analysis, the types of cyberbullying were sorted using SPSS 25.0. The results indicated that the total rate of types of cyberbullying in Korea was 44%, which consisted of 32.3% victims, 6.4% perpetrators, and 5.3% bystanders. According to the results, the impulse factor was also the greatest influence on the prediction of the risk factors and the propensity for dominance factor was the second greatest factor predicting the types of risk factors. In particular, the impulse factor had the most significant effect on bystanders, and the propensity for dominance factor was also significant in influencing online perpetrators. It is necessary to develop a program to diminish the impulses that were initiated by bystanders as well as victims and perpetrators because many of those bystanders have tended to aggravate impulsive cyberbullying behaviors.     

A pattern recognition-based approach for phylogenetic network construction with constrained recombination

The tree representation of evolutionary relationship oversimplifies the view of the process of evolution as it cannot take into account the events such as horizontal gene transfer, hybridization, homoplasy and genetic recombination. Several algorithms exist for constructing phylogenetic networks which result from events such as horizontal gene transfer, hybridization and homoplasy. Very little work has been published on the algorithmic detail of phylogenetic networks with constrained recombination. The problem of minimizing the number of recombinations in a phylogenetic network, constructed using binary DNA sequences, is NP-hard. In this paper, we propose a pattern recognition-based O(n²) time approach for constructing the phylogenetic network, where n is the number of nodes or sequences in the input data. The network is constructed with the restriction that no two cycles in the network share a common node.     

On finding optimal and near-optimal lineal spanning trees

The search for good lineal, or depth-first, spanning trees is an important aspect in the implementation of a wide assortment of graph algorithms. We consider the complexity of findingoptimal lineal spanning trees under various notions of optimality. In particular, we show that several natural problems, such as constructing a shortest or a tallest lineal tree, are NP-hard. We also address the issue of polynomial-time, near-optimization strategies for these difficult problems, showing that efficient absolute approximation algorithms cannot exist unlessP = NP.This author's research was supported in part by the Sandia University Research Program and by the National Science Foundation under Grant M IP-8603879.This author's research was supported in part by the National Science Foundation under Grants ECS-8403859 and MIP-8603879.     

On the cost of searching signature trees

A precise analysis of the retrieval of signature trees is presented. A signature tree is a data structure constructed over a signature file to speed up searching all those signatures, which match a given query signature. The methods used include a detailed study of probabilistic analysis in conjunction with suitable contour integration of complex variabled functions.     

Technical Note: Selecting a Classification Method by Cross-Validation

If we lack relevant problem-specific knowledge, cross-validation methods may be used to select a classification method empirically. We examine this idea here to show in what senses cross-validation does and does not solve the selection problem. As illustrated empirically, cross-validation may lead to higher average performance than application of any single classification strategy, and it also cuts the risk of poor performance. On the other hand, cross-validation is no more or less a form of bias than simpler strategies, and applying it appropriately ultimately depends in the same way on prior knowledge. In fact, cross-validation may be seen as a way of applying partial information about the applicability of alternative classification strategies.     

Improved wetland remote sensing in Yellowstone National Park using classification trees to combine TM imagery and ancillary environmental data

The U.S. Fish and Wildlife Service uses the term palustrine wetland to describe vegetated wetlands traditionally identified as marsh, bog, fen, swamp, or wet meadow. Landsat TM imagery was combined with image texture and ancillary environmental data to model probabilities of palustrine wetland occurrence in Yellowstone National Park using classification trees. Model training and test locations were identified from National Wetlands Inventory maps, and classification trees were built for seven years spanning a range of annual precipitation. At a coarse level, palustrine wetland was separated from upland. At a finer level, five palustrine wetland types were discriminated: aquatic bed (PAB), emergent (PEM), forested (PFO), scrub-shrub (PSS), and unconsolidated shore (PUS). TM-derived variables alone were relatively accurate at separating wetland from upland, but model error rates dropped incrementally as image texture, DEM-derived terrain variables, and other ancillary GIS layers were added. For classification trees making use of all available predictors, average overall test error rates were 7.8% for palustrine wetland/upland models and 17.0% for palustrine wetland type models, with consistent accuracies across years. However, models were prone to wetland over-prediction. While the predominant PEM class was classified with omission and commission error rates less than 14%, we had difficulty identifying the PAB and PSS classes. Ancillary vegetation information greatly improved PSS classification and moderately improved PFO discrimination. Association with geothermal areas distinguished PUS wetlands. Wetland over-prediction was exacerbated by class imbalance in likely combination with spatial and spectral limitations of the TM sensor. Wetland probability surfaces may be more informative than hard classification, and appear to respond to climate-driven wetland variability. The developed method is portable, relatively easy to implement, and should be applicable in other settings and over larger extents.     

An empirical InSAR-optical fusion approach to mapping vegetation canopy height

Exploiting synergies afforded by a host of recently available national-scale data sets derived from interferometric synthetic aperture radar (InSAR) and passive optical remote sensing, this paper describes the development of a novel empirical approach for the provision of regional- to continental-scale estimates of vegetation canopy height. Supported by data from the 2000 Shuttle Radar Topography Mission (SRTM), the National Elevation Dataset (NED), the LANDFIRE project, and the National Land Cover Database (NLCD) 2001, this paper describes a data fusion and modeling strategy for developing the first-ever high-resolution map of canopy height for the conterminous U.S. The approach was tested as part of a prototype study spanning some 62,000 km² in central Utah (NLCD mapping zone 16). A mapping strategy based on object-oriented image analysis and tree-based regression techniques is employed. Empirical model development is driven by a database of height metrics obtained from an extensive field plot network administered by the USDA Forest Service-Forest Inventory and Analysis (FIA) program. Based on data from 508 FIA field plots, an average absolute height error of 2.1 m (r = 0.88) was achieved for the prototype mapping zone.     

Using airborne laser scanning to monitor tree migration in the boreal-alpine transition zone

The boreal tree line is expected to advance upwards into the mountains and northwards into the tundra due to global warming. The major objective of this study was to find out if it is possible to use high-resolution airborne laser scanner data to detect very small trees — the pioneers that are pushing the tree line up into the mountains and out onto the tundra. The study was conducted in a sub-alpine/alpine environment in southeast Norway. A total of 342 small trees of Norway spruce, Scots pine, and downy birch with tree heights ranging from 0.11 to 5.20 m were precisely georeferenced and measured in field. Laser data were collected with a pulse density of 7.7 m^− 2. Three different terrain models were used to process the airborne laser point cloud in order to assess the effects of different pre-processing parameters on small tree detection. Greater than 91% of all trees > 1 m tall registered positive laser height values regardless of terrain model. For smaller trees (< 1 m), positive height values were found in 5-73% of the cases, depending on the terrain model considered. For this group of trees, the highest rate of trees with positive height values was found for spruce. The more smoothed the terrain model was, the larger the portion of the trees that had positive laser height values. The accuracy of tree height derived from the laser data indicated a systematic underestimation of true tree height by 0.40 to 1.01 m. The standard deviation for the differences between laser-derived and field-measured tree heights was 0.11-0.73 m. Commission errors, i.e., the detection of terrain objects — rocks, hummocks — as trees, increased significantly as terrain smoothing increased. Thus, if no classification of objects into classes like small trees and terrain objects is possible, many non-tree objects with a positive height value cannot be separated from those actually being trees. In a monitoring context, i.e., repeated measurements over time, we argue that most other objects like terrain structures, rocks, and hummocks will remain stable over time while the trees will change as they grow and new trees are established. Thus, this study indicates that, given a high laser pulse density and a certain density of newly established trees, it would be possible to detect a sufficient portion of newly established trees over a 10 years period to claim that tree migration is taking place.