Conventional Statistical Techniques Research Articles

Nonlinearity in the Indian commodity markets: evidence from a battery of tests

This study tests for the presence of nonlinear dependence in the rate of returns series for four Indian multi commodity exchange indices. Six different tests for detecting nonlinearity were employed and the statistics were estimated using both the asymptotic theory and the bootstrap. The analysis of results reveals that there is a consensus in favour of nonlinearity for Metal and Energy indices even after removing linear serial correlations from the data, hence, contradicting the unpredictable criterion of weak-form efficient market hypothesis. Further the results from additional BDS test on the standardised AR(P)-GARCH(1,1) residuals imply that conditional heteroskedasticity is the main source of nonlinearity in metal indices and could be captured by the ARCH-type models. On the other hand in case of energy indices results indicate non-linear dependence of an unknown form in the data. The results have important implications on the earlier work on Indian commodity derivative markets which have relied on conventional statistical techniques which may have lead to biased and highly misleading results.

Association Rule Mining in Developmental Psychology

Developmental psychology is the scientific study of progressive psychological changes that occur in human beings as they age. Some of the current methodologies used in this field to study developmental processes include Yule’s Q, state space grids, time series analysis, and lag analysis. The data collected in this field are often time-series-type data. Applying association rule mining in developmental psychology is a new concept that may have a number of potential benefits. In this paper, two sets of infant-mother interaction data sets are examined using association rule mining. Previous analyses of these data used conventional statistical techniques. However, they failed to capture the dynamic interactions between the infant-mother pair as well as other issues relating to the temporal characteristic of the data. Three approaches are proposed in this paper as candidate means of addressing some of the questions that remain from previous studies. The approaches used can be applied to association rule mining to extend its application to data sets in related fields.

A New Hybrid Method Logistic Regression and Feedforward Neural Network for Lung Cancer Data

Logistic regression (LR) is a conventional statistical technique used for data classification problem. Logistic regression is a model‐based method, and it uses nonlinear model structure. Another technique used for classification is feedforward artificial neural networks. Feedforward artificial neural network is a data‐based method which can model nonlinear models through its activation function. In this study, a hybrid approach of model‐based logistic regression technique and data‐based artificial neural network was proposed for classification purposes. The proposed approach was applied to lung cancer data, and obtained results were compared. It was seen that the proposed hybrid approach was superior to logistic regression and feedforward artificial neural networks with respect to many criteria.

Analysis of Data From the Barnett Shale Using Conventional Statistical and Virtual Intelligence Techniques

Summary A Barnett shale water-production data set from approximately 11,000 completions was analyzed using conventional statistical techniques. Additionally, a water/hydrocarbon ratio and first-derivative diagnostic-plot technique developed elsewhere for conventional reservoirs was extended to analyze Barnett shale water-production mechanisms. To determine hidden structure in well and production data, self-organizing maps and the k-means algorithm were used to identify clusters in data. A competitive-learning-based network was used to predict the potential for continuous water production from a new well, and a feed-forward neural network was used to predict average water production for wells drilled in Denton and Parker Counties, Texas, of the Barnett shale. Using conventional techniques, we concluded that for wells of the same completion type, location is more important than time of completion or hydraulic-fracturing strategy. Liquid loading has potential to affect vertical more than horizontal wells. Different features were observed in the spreadsheet diagnostic plots for wells in the Barnett shale, and we made a subjective interpretation of these features. We find that 15% of the horizontal and vertical wells drilled in Denton County have a load-water-recovery factor greater than unity. Also, 15 and 35% of the horizontal and vertical wells drilled, respectively, in Parker County have a load-recovery factor greater than unity.The use of both self-organizing maps and the k-means algorithm showed that the data set is divided into two main clusters. The physical properties of these clusters are unknown but interpreted to represent wells with high water throughput and those with low water throughput. Expected misclassification error for the competitive-learning-based tool was approximately 10% for a data set containing both vertical and horizontal wells. The average prediction error for the neural-network tool varied between 10 and 26%, depending on well type and location.Results from this work can be used to mitigate risk of water problems in new Barnett shale wells and predict water issues in other shale plays. Engineers are provided a tool to predict potential for water production in new wells. The method used to develop this tool can be used to solve similar challenges in new and existing shale plays.

Bayesian networks for defining relationships among climate factors

Factors that shape the weather are studied to design models to make predictions. Understanding the relationships among these factors contribute to a better knowledge of atmospheric phenomena. However, conventional statistical techniques do not take into account the dependent relationships among these factors. Bayesian learning models are used in learning processes, which consist in quantification of conditional probability, resulting in the identification of causal relationships between the variables. In this paper the use of Bayesian networks for probabilistic analysis allows us to determine spatial and temporal dependencies among climatic variables not observable with other methods. Considering an incomplete meteorological data set from three years and three sites with distinct climates, the dependent relationships between climatic variables are observable in different proportions for each type of climate. It is possible to determine the influences among variables, temperature, humidity, dew point, pressure, wind speed and precipitation, through the use of Bayesian networks that permit us to understand their interaction in different climates. Key words: Probability models, climate, forecast, data mining, K2 algorithm.

A Fuzzy Approach For Clustering Gene Expression Time Series Data

Identifying groups of genes that manifest similar expression patterns is crucial in the analysis of gene expression time series data. Choosing a similarity measure to determine the similarity or distance between profiles is an important task. Time series expression experiments are used to study a wide range of biological systems. More than 80% of all time series expression datasets are short (8 time points or fewer). These datasets present unique challenges. On account of the large number of genes profiled (often tens of thousands) and the small number of time points many patterns are expected to arise at random. Most clustering algorithms are unable to distinguish between real and random patterns. However, the shortness of gene expression time-series data limits the use of conventional statistical models and techniques for time-series analysis. To address this problem, this paper proposes the Fuzzy clustering algorithm based on short time-series, which is able to cluster profiles based on the similarity of their relative change of expression level and the corresponding temporal in- formation. One of the major advantages of fuzzy clustering is that genes can belong to more than one group, revealing distinctive features of each gene's function and regulation.

Self-Organizing Feature Maps and selected conventional numerical methods for assessment of environmental quality

The investigations concerned sites of <em>Acer platanoides</em> L. infected or not by <em>Rhytisma aceriniu</em> (Pers.) Fr. The aim of the study was to check the occurrence of <em>R. acerinium</em>, and whether it reflects the environmental status. Furthermore, an analysis was carried out to find out whether the applied SOFM offers additional advantages to solve problems in relation to conventional methods. Concentrations of selected elements in soils and leaves, and leaf and "tar-spot" morphometric traits were also measured. A significant differentiation was found between sites in relation to the analyzed traits. It appeared, that sites showing lower concentrations of chemical elements and proper developmental habitat conditions massive infections take place. The study showed that <em>R. acerinium</em> is a good biological indicator for assessment of environmental status. The applied, conventional statistical methods, SOFM and image techniques showed similar, but not identical results for assessment of environmental quality using <em>R. acerinium</em>. SOFM appeared to be more useful for ordination of results and ought to be taken into account as a proper tool of estimation of various plants and their biotopes.

Comparing techniques for vegetation classification using multi- and hyperspectral images and ancillary environmental data

This paper evaluates the predictive power of innovative and more conventional statistical classification techniques. We use Landsat 7 Enhanced Thematic Mapper Plus (ETM+), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and airborne imaging spectrometer (HyMap) images to classify Mediterranean vegetation types, with and without inclusion of ancillary data (geology, soil classes and digital elevation model derivatives). When the number of classes is low, both conventional and innovative techniques perform well. For larger numbers of classes the innovative techniques of random forests and support vector machines outperform the other techniques. Compared to conventional techniques, classification trees, random forests and support vector machines proved to be better suited for the incorporation of continuous and categorical ancillary data: overall accuracies and accuracies for individual classes improve significantly when many, difficult to separate, classes are taken into account. Therefore, these techniques are definitely worth including in common image analysis software packages.

Technology Focus: Gas Production Technology (November 2010)

Technology Focus Those who cannot remember the past are condemned to repeat it. George Santayana Rarely a day goes by when a new shale-gas conference is not advertised; each will highlight the rapidly changing state of the art in fractured horizontal wells and shed light on what will always be a very complex and challenging way to complete wells. With increased interest and activity comes the need to evaluate and forecast the performance of these wells accurately with little production history as a basis. This problem has plagued engineers since the first wells were drilled in the 19th century. How does one predict the future of a new kind of well when few examples exist with enough production history to guide the forecast, and the reservoir behaviors are not understood well enough to simulate the performance mathematically? Lessons from the past may help. Typically, a new play is ignited by an operator that is stubborn enough to keep trying new things while ignoring years of failure and the peer consensus that the operator is crazy. Soon after the secret is discovered, this newly exposed fertile ground attracts other operators desperate for better opportunities. The increased demand for leases, equipment, and services and increased regulation drive the economic returns down, not to mention the glut of production if the play lives up to its promise. Meanwhile, a new set of contrarians will begin to question the assumptions that created the boom in the first place; because rarely is there enough history to definitively prove the long-term outcomes before most have lost interest and moved on to the next hot play. Buffering this emotional ebb and flow are operators who are not swayed by the prevailing thought. The selections for this month’s feature are examples from those who diligently look for patterns from the past to improve their accuracy in predicting the future. Gas Production Technology additional reading available at OnePetro: www.onepetro.org SPE 127919 • “Analysis of Data From the Barnett Shale Using Conventional Statistical and Virtual-Intelligence Techniques,” by Obadare O. Awoleke, Texas A&M University, et al. SPE 130151 • “Enhanced Gas Recovery—Challenges Shown in the Example of Three Gas Fields” by Torsten Clemens, OMV, et al. SPE 131714 • “The Permeability Factor in Coalbed-Methane-Well Completions and Production” by Ian Palmer, Higgs-Palmer Technologies

Prediction of daily rainfall state in a river basin using statistical downscaling from GCM output

Conventional statistical downscaling techniques for prediction of multi-site rainfall in a river basin fail to capture the correlation between multiple sites and thus are inadequate to model the variability of rainfall. The present study addresses this problem through representation of the pattern of multi-site rainfall using rainfall state in a river basin. A model based on K-means clustering technique coupled with a supervised data classification technique, namely Classification And Regression Tree (CART), is used for generation of rainfall states from large-scale atmospheric variables in a river basin. The K-means clustering is used to derive the daily rainfall state from the historical daily multi-site rainfall data. The optimum number of clusters in the observed rainfall data is obtained after application of various cluster validity measures to the clustered data. The CART model is then trained to establish relationship between the daily rainfall state of the river basin and the standardized, dimensionally-reduced National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis climatic data set. The relationship thus developed is applied to the General Circulation Model (GCM)-simulated, standardized, bias free large-scale climate variables for prediction of rainfall states in future. Comparisons of the number of days falling under different rainfall states for the observed period and the future give the change expected in the river basin due to global warming. The methodology is tested for the Mahanadi river basin in India.

Intelligent Diagnosis of Heart Diseases using Neural Network Approach

Experiments with the Switzerland Heart Disease database have concentrated on attempting to distinguish presence and absence. The classifiers based on various neural networks, namely, MLP, PCA, Jordan, GFF, Modular, RBF, SOFM, SVM NNs and conventional statistical techniques such as DA and CART are optimally designed, thoroughly examined and performance measures are compared in this study. With chosen optimal parameters of MLP NN, when it is trained and tested over cross validation (unseen data sets), the average (and best respectively) classification of 98±2.83 % (and 100%), 96.67±4.56% overall accuracy, sensitivity 96±5.48, specificity 100% are achieved which shows consistent performance than other NN and statistical models. The results obtained in this work show the potentiality of the MLP NN approach for heart diseases classification.

Particle and Microorganism Enumeration Data: Enabling Quantitative Rigor and Judicious Interpretation

Many of the methods routinely used to quantify microscopic discrete particles and microorganisms are based on enumeration, yet these methods are often known to yield highly variable results. This variability arises from sampling error and variations in analytical recovery (i.e., losses during sample processing and errors in counting), and leads to considerable uncertainty in particle concentration or log(10)-reduction estimates. Conventional statistical analysis techniques based on the t-distribution are often inappropriate, however, because the data must be corrected for mean analytical recovery and may not be normally distributed with equal variance. Furthermore, these statistical approaches do not include subjective knowledge about the stochastic processes involved in enumeration. Here we develop two probabilistic models to account for the random errors in enumeration data, with emphasis on sampling error assumptions, nonconstant analytical recovery, and discussion of counting errors. These models are implemented using Bayes' theorem to yield posterior distributions (by numerical integration or Gibbs sampling) that completely quantify the uncertainty in particle concentration or log(10)-reduction given the experimental data and parameters that describe variability in analytical recovery. The presented approach can easily be implemented to correctly and rigorously analyze single or replicate (bio)particle enumeration data.

New approaches to modelling fish–habitat relationships

Ecologists often develop models that describe the relationship between faunal communities and their habitat. Coral reef fishes have been the focus of numerous such studies, which have used a wide range of statistical tools to answer an equally wide range of questions. Here, we apply a series of both conventional statistical techniques (linear and generalized additive regression models) and novel machine-learning techniques (the support vector machine and three ensemble techniques used with regression trees) to predict fish species richness, biomass, and diversity from a range of habitat variables. We compare the techniques in terms of their predictive performance, and we compare a subset of the models in terms of the influence each habitat variable has for the predictions. Prediction errors are estimated by cross-validation, and variable importance is assessed using permutations of individual variable values. For predictions of species richness and diversity the tree-based models generally and the random forest model specifically are superior (produce the lowest errors). These model types are all able to model both nonlinear and interaction effects. The linear model, unable to model either effect type, performs the worst (produces the highest errors). For predictions of biomass, the generalized additive model is superior, and the support vector machine performs the worst. Depth range, the difference between maximum and minimum water depth at a given site, is identified as the most important variable in the majority of models predicting the three fish community variables. However, variable importance is highly dependent upon model type, which leads to questions regarding the interpretation of variable importance and its proper use as an indicator of causality. The representation of ecological relationships by tree-based ensemble learners will improve predictive performance, and provide a new avenue for exploring ecological relationships, both statistical and causal.

Reservoir properties prediction in the West Baram Delta through data integration constrained by rock physics

We present the methodology and results that allowed conditioning the geological model in Siwa Field in Malaysia's Sarawak Basin using seismic-derived properties. Synergic integration of advanced petrophysical techniques, rock physics analysis/modeling, and seismic inversion, combined with a detailed structural modeling and depth conversion, allowed us to successfully incorporate the seismic-driven petrophysical properties and achieve a more complete description of the reservoir than could be done through conventional statistical techniques based on wells and seismic maps. The high structural and stratigraphic complexity, lack of shear-wave velocity information, and high levels of noise in the data represented the biggest challenges to deriving seismic properties based on amplitude variation for conditioning the reservoir model.

Introduction to the DISRUPT postprandial database: subjects, studies and methodologies

Dysregulation of lipid and glucose metabolism in the postprandial state are recognised as important risk factors for the development of cardiovascular disease and type 2 diabetes. Our objective was to create a comprehensive, standardised database of postprandial studies to provide insights into the physiological factors that influence postprandial lipid and glucose responses. Data were collated from subjects (n = 467) taking part in single and sequential meal postprandial studies conducted by researchers at the University of Reading, to form the DISRUPT (DIetary Studies: Reading Unilever Postprandial Trials) database. Subject attributes including age, gender, genotype, menopausal status, body mass index, blood pressure and a fasting biochemical profile, together with postprandial measurements of triacylglycerol (TAG), non-esterified fatty acids, glucose, insulin and TAG-rich lipoprotein composition are recorded. A particular strength of the studies is the frequency of blood sampling, with on average 10-13 blood samples taken during each postprandial assessment, and the fact that identical test meal protocols were used in a number of studies, allowing pooling of data to increase statistical power. The DISRUPT database is the most comprehensive postprandial metabolism database that exists worldwide and preliminary analysis of the pooled sequential meal postprandial dataset has revealed both confirmatory and novel observations with respect to the impact of gender and age on the postprandial TAG response. Further analysis of the dataset using conventional statistical techniques along with integrated mathematical models and clustering analysis will provide a unique opportunity to greatly expand current knowledge of the aetiology of inter-individual variability in postprandial lipid and glucose responses.

Artificial Bee Colony for Classification of Acoustic Emission Signal Source

In this paper the approach for classification of acoustic emission signals to their respective sources is employed using a swarm intelligence technique called artificial bee colony. In this work, artificial bee colony is employed to train a multilayer perceptron neural network which is used for the classification of the acoustic emission signal to their respective source. Acoustic emission is carried out using pulse, pencil and spark signal source on the surface of solid steel block. The signal parameters are measured using AET 5000 system. To begin with, the complexity for acoustic emission data set is verified using conventional statistical technique like principal component analysis and traditional training algorithm like multilayer perceptron neural network trained using the backpropagation algorithm. The experiment shows in both the case the classification is not accurate. For this complex acoustic emission data set multilayer perceptron neural network trained using the artificial bee colony algorith...

Predicting the graft survival for heart–lung transplantation patients: An integrated data mining methodology

Background Predicting the survival of heart–lung transplant patients has the potential to play a critical role in understanding and improving the matching procedure between the recipient and graft. Although voluminous data related to the transplantation procedures is being collected and stored, only a small subset of the predictive factors has been used in modeling heart–lung transplantation outcomes. The previous studies have mainly focused on applying statistical techniques to a small set of factors selected by the domain-experts in order to reveal the simple linear relationships between the factors and survival. The collection of methods known as ‘data mining’ offers significant advantages over conventional statistical techniques in dealing with the latter's limitations such as normality assumption of observations, independence of observations from each other, and linearity of the relationship between the observations and the output measure(s). There are statistical methods that overcome these limitations. Yet, they are computationally more expensive and do not provide fast and flexible solutions as do data mining techniques in large datasets. Purpose The main objective of this study is to improve the prediction of outcomes following combined heart–lung transplantation by proposing an integrated data-mining methodology. Methods A large and feature-rich dataset (16,604 cases with 283 variables) is used to (1) develop machine learning based predictive models and (2) extract the most important predictive factors. Then, using three different variable selection methods, namely, (i) machine learning methods driven variables—using decision trees, neural networks, logistic regression, (ii) the literature review-based expert-defined variables, and (iii) common sense-based interaction variables, a consolidated set of factors is generated and used to develop Cox regression models for heart–lung graft survival. Results The predictive models’ performance in terms of 10-fold cross-validation accuracy rates for two multi-imputed datasets ranged from 79% to 86% for neural networks, from 78% to 86% for logistic regression, and from 71% to 79% for decision trees. The results indicate that the proposed integrated data mining methodology using Cox hazard models better predicted the graft survival with different variables than the conventional approaches commonly used in the literature. This result is validated by the comparison of the corresponding Gains charts for our proposed methodology and the literature review based Cox results, and by the comparison of Akaike information criteria (AIC) values received from each. Conclusions Data mining-based methodology proposed in this study reveals that there are undiscovered relationships (i.e. interactions of the existing variables) among the survival-related variables, which helps better predict the survival of the heart–lung transplants. It also brings a different set of variables into the scene to be evaluated by the domain-experts and be considered prior to the organ transplantation.

Nonlinear association between serum testosterone levels and coronary artery disease in Iranian men

Previous studies have shown controversial results about the role of androgens in coronary artery disease (CAD). We performed this study to examine and compare the relationship between androgenic hormones and CAD using conventional linear statistical techniques as well as novel non-linear approaches. The study was conducted on 502 consecutive men who were referred for selective coronary angiography at Tehran Heart Center due to different indications. We studied the relationship between androgenic hormones and CAD by using the generalized linear models, generalized additive models, and neural networks. Free testosterone (fT), total testosterone (tT) and dehydroepiandrosterone sulfate levels in patients with significant CAD versus normal individuals were 6.69 +/- 3.20 pg/ml, 16.60 +/- 6.66 nm/l, and 113.38 +/- 72.9 microg/dl versus 7.12 +/- 3.58 pg/ml, 15.82 +/- 7.26 nm/l, and 109.03 +/- 68.19 microg/dl, respectively (P > 0.05). The Generalized linear models was unable to show any significant relationship between androgenic hormones and CAD, while generalized additive model and neural networks supported the significant effect of androgenic hormones on CAD. This finding suggests a nonlinear association of tT levels with CAD: lower levels have a preventive effect on CAD, whereas higher values increase the risk of CAD. Emphasizing the non-linearity of the variables may provide new insight into the possible explanation of the effect of androgenic hormones on CAD.

Maximization of representativity factors for experimental planning of cross-section measurements: An information theoretic approach

The similarity between two nuclear systems is expressed by the representativity factors and maximizing its value to unity reduces the uncertainty from its existing value. As the representativity factor is a function of covariance matrix of neutron cross-sections, the inherent systematic uncertainty in neutron cross-sections inhibits it from approaching unity and hence statistical procedures have to be resorted to minimize the systematic uncertainty. As the conventional statistical techniques fail when systematic uncertainty is dominant, we propose an entropy based information theoretic approach of maximizing the mutual information by the knowledge of bounds for the correlated elements. We show that maximizing the mutual information and the representativity factors express the similar phenomena of uncertainty reduction. We estimated the bounds for the correlated elements of the correlation matrix for minor actinides and show how the systematic uncertainty is reduced when lower bound values are considered. These lower bound values aid in experimental planning for future measurement of cross-sections with reduced systematic uncertainty.

Rule extraction for classification of acoustic emission signals using Ant Colony Optimisation

Ant Colony Optimization (ACO) is used to obtain rules that can classify the data into pre-defined classes. It can be used to classify acoustic emission (AE) signals to their respective sources. ACO based technique has an advantage over conventional statistical techniques like maximum likelihood estimate, nearest neighbor classifier, etc., because they are distribution free, i.e., no knowledge is required about the distribution of data. AE test is carried out using pulse, pencil and spark signal source on the surface of solid steel block. The signal parameters are measured using AET 5000 system. Classification of AE signal is done using Ant Colony Optimization, and the simplicity of the rules generated is emphasized.