Improve Prediction Quality Research Articles

Integrated cluster analysis and artificial neural network modeling for steam-assisted gravity drainage performance prediction in heterogeneous reservoirs

Evaluation of steam-assisted gravity drainage (SAGD) performance that involves detailed compositional simulations is usually deterministic, cumbersome, expensive (manpower and time consuming), and not quite suitable for practical decision making and forecasting, particularly when dealing with high-dimensional data space consisting of large number of operational and geological parameters. Data-driven modeling techniques, which entail comprehensive data analysis and implementation of machine learning methods for system forecast, provide an attractive alternative.In this paper, artificial neural network (ANN) is employed to predict SAGD production in heterogeneous reservoirs, an important application that is lacking in existing literature. Numerical flow simulations are performed to construct a training data set consists of various attributes describing characteristics associated with reservoir heterogeneities and other relevant operating parameters. Empirical Arps decline parameters are tested successfully for parameterization of cumulative production profile and considered as outputs of the ANN models. Sensitivity studies on network configurations are also investigated. Principal components analysis (PCA) is performed to reduce the dimensionality of the input vector, improve prediction quality, and limit over-fitting. In a case study, reservoirs with distinct heterogeneity distributions are fed to the model. It is shown that robustness and accuracy of the prediction capability are greatly enhanced when cluster analysis are performed to identify internal data structures and groupings prior to ANN modeling. Both deterministic and fuzzy-based clustering techniques are compared, and separate ANN model is constructed for each cluster. The model is then tested using a validation data set (cases that have not been used during the training stage).The proposed approach can be integrated directly into most existing reservoir management routines. In addition, incorporating techniques for dimensionality reduction and clustering with ANN demonstrates the viability of this approach for analyzing large field data set. Given that quantitative ranking of operating areas, robust forecasting, and optimization of heavy oil recovery processes are major challenges faced by the industry, the proposed research highlights the significant potential of applying effective data-driven modeling approaches in analyzing other solvent-additive steam injection projects.

Comparative analysis of local and consensus quantitative structure-activity relationship approaches for the prediction of bioconcentration factor

Quantitative structure-activity relationships (QSARs) are broadly classified as global or local, depending on their molecular constitution. Global models use large and diverse training sets covering a wide range of chemical space. Local models focus on smaller structurally or chemically similar subsets that are conventionally selected by human experts or alternatively using clustering analysis. The current study focuses on the comparative analysis of different clustering algorithms (expectation-maximization, K-means and hierarchical) for seven different descriptor sets as structural characteristics and two rule-based approaches to select subsets for designing local QSAR models. A total of 111 local QSAR models are developed for predicting bioconcentration factor. Predictions from local models were compared with corresponding predictions from the global model. The comparison of coefficients of determination (r 2) and standard deviations for local models with similar subsets from the global model show improved prediction quality in 97% of cases. The descriptor content of derived QSARs is discussed and analyzed. Local QSAR models were further consolidated within the framework of consensus approach. All different consensus approaches increased performance over the global and local models. The consensus approach reduced the number of strongly deviating predictions by evening out prediction errors, which were produced by some local QSARs.

Local Practice of Assessing Dynamic Modulus Properties for Washington State Mixtures

The dynamic modulus (|E*|) is one of the key elements of a mechanistic–empirical–based flexible pavement design procedure. The dynamic modulus is used to characterize the material properties of asphalt mixtures and to determine the stress–strain responses of a pavement at different loading conditions. The dynamic modulus is also a direct input parameter in several pavement performance models to estimate field fatigue cracking and rutting performance. To provide a better understanding of the local materials, this study aimed to test the typical asphalt mixtures used by the Washington State Department of Transportation and to establish a material catalog for dynamic modulus. In this study, seven plant-produced mixtures from five regions of Washington State were sampled and tested. These mixtures represented the typical asphalt binder, gradation, and mix designs of the state. One warm-mix asphalt project was also included in the analysis. On the basis of the experimental results, it was found that mix properties including air voids and binder properties had an important impact on the dynamic modulus. Because of the limited aggregate gradations, the effect of aggregate gradation on the dynamic modulus was inconclusive. The measured dynamic modulus data were compared with the prediction results by using the traditional Witczak E* model, the new Witczak E* model, and the Hirsch model. The Hirsch model was found to be the most promising and was modified further by including mastic property into the model. The modified Hirsch model greatly improved prediction quality and can be used as both a design tool and a screening tool to estimate the dynamic modulus of a mixture at early stages of the mix design process.

Multiple information contents derived from the chromatograms and their application to the modeling of quantitative profile–efficacy relationship

Herbal medicine has been successfully applied in clinical therapeutics throughout the world. In this paper, various information contents in addition to a best-first search strategy are proposed for improving prediction quality of quantitative profile–efficacy relationship (QPER). As a demonstration, a QPER model was constructed to predict the blood-clotting time increasing effect of Naodesheng prescription, a widely used herbal medicine in China. Evaluated by both the self-consistency test and the rigorous jackknife test, the predictive accuracy was increased by feeding the model with the merged information contents compared to only with the common chromatographic peak areas. The confirmation experiment exhibited that the predicted blood-clotting time were well consistent with the observed values. Furthermore, it has not escaped from our notice that the important potential application of the current QPER model is to be extended to help search active fractions, which has attracted more and more attention in the re-developments of herbal medicines based on the active compounds.

Rich Parameterization Improves RNA Structure Prediction

Current approaches to RNA structure prediction range from physics-based methods, which rely on thousands of experimentally measured thermodynamic parameters, to machine-learning (ML) techniques. While the methods for parameter estimation are successfully shifting toward ML-based approaches, the model parameterizations so far remained fairly constant. We study the potential contribution of increasing the amount of information utilized by RNA folding prediction models to the improvement of their prediction quality. This is achieved by proposing novel models, which refine previous ones by examining more types of structural elements, and larger sequential contexts for these elements. Our proposed fine-grained models are made practical thanks to the availability of large training sets, advances in machine-learning, and recent accelerations to RNA folding algorithms. We show that the application of more detailed models indeed improves prediction quality, while the corresponding running time of the folding algorithm remains fast. An additional important outcome of this experiment is a new RNA folding prediction model (coupled with a freely available implementation), which results in a significantly higher prediction quality than that of previous models. This final model has about 70,000 free parameters, several orders of magnitude more than previous models. Being trained and tested over the same comprehensive data sets, our model achieves a score of 84% according to the F₁-measure over correctly-predicted base-pairs (i.e., 16% error rate), compared to the previously best reported score of 70% (i.e., 30% error rate). That is, the new model yields an error reduction of about 50%. Trained models and source code are available at www.cs.bgu.ac.il/?negevcb/contextfold.

An Effective Hybrid Semi-Parametric Regression Strategy for Rainfall Forecasting Combining Linear and Nonlinear Regression

Rainfall forecasting is an important research topic in disaster prevention and reduction. The characteristic of rainfall involves a rather complex systematic dynamics under the influence of different meteorological factors, including linear and nonlinear pattern. Recently, many approaches to improve forecasting accuracy have been introduced. Artificial neural network (ANN), which performs a nonlinear mapping between inputs and outputs, has played a crucial role in forecasting rainfall data. In this paper, an effective hybrid semi-parametric regression ensemble (SRE) model is presented for rainfall forecasting. In this model, three linear regression models are used to capture rainfall linear characteristics and three nonlinear regression models based on ANN are able to capture rainfall nonlinear characteristics. The semi-parametric regression is used for ensemble model based on the principal component analysis technique. Empirical results reveal that the prediction using the SRE model is generally better than those obtained using other models in terms of the same evaluation measurements. The SRE model proposed in this paper can be used as a promising alternative forecasting tool for rainfall to achieve greater forecasting accuracy and improve prediction quality.

Blind, one‐eyed, or eagle‐eyed? pKa calculations during blind predictions with staphylococcal nuclease

In the current contribution, the performance of Poisson-Boltzmann-based pK(a) calculations of SNase mutants as part of a blind prediction exercise facilitated by the pK(a) cooperative ("pK(a) _coop") is described. A one parameter setting ("quick&dirty" approach) is used to provide an industry perspective where strong time constraints are frequently encountered. On the one hand, results are analyzed in terms of root mean square deviation performance. Furthermore, the pK(a) calculations are assessed for their ability to properly assign protonation state. For this purpose, a new measure called BIPS (binary protonation state at physiological pH) is introduced. Significant differences were found with both comparison measures based on the class of residues examined. In addition, the performance of PROPKA3 as well as the NULL model is examined on the same data set. Finally, pK(a) calculations on SNase mutants with available structural information have been performed and provide support for our calculation methods. The performance on this subset is better than on the pK(a) cooperative mutation data. In the pK(a) _coop data, no structural information on the generated mutants is available. This suggests the occurrence of a substantial structural rearrangement on the insertion of additional charged groups into SNase, which leads to improved prediction quality.

Reducing the worst case running times of a family of RNA and CFG problems, using Valiant’s approach

BackgroundRNA secondary structure prediction is a mainstream bioinformatic domain, and is key to computational analysis of functional RNA. In more than 30 years, much research has been devoted to defining different variants of RNA structure prediction problems, and to developing techniques for improving prediction quality. Nevertheless, most of the algorithms in this field follow a similar dynamic programming approach as that presented by Nussinov and Jacobson in the late 70's, which typically yields cubic worst case running time algorithms. Recently, some algorithmic approaches were applied to improve the complexity of these algorithms, motivated by new discoveries in the RNA domain and by the need to efficiently analyze the increasing amount of accumulated genome-wide data.ResultsWe study Valiant's classical algorithm for Context Free Grammar recognition in sub-cubic time, and extract features that are common to problems on which Valiant's approach can be applied. Based on this, we describe several problem templates, and formulate generic algorithms that use Valiant's technique and can be applied to all problems which abide by these templates, including many problems within the world of RNA Secondary Structures and Context Free Grammars.ConclusionsThe algorithms presented in this paper improve the theoretical asymptotic worst case running time bounds for a large family of important problems. It is also possible that the suggested techniques could be applied to yield a practical speedup for these problems. For some of the problems (such as computing the RNA partition function and base-pair binding probabilities), the presented techniques are the only ones which are currently known for reducing the asymptotic running time bounds of the standard algorithms.

Self-similarity: Behind workload reshaping and prediction

Energy efficiency has become one of the most important challenges in designing future computing systems. Workload reshaping and prediction have a significant impact on the energy conservation of computer components. In this paper, by using three HTTP traffic traces, we perform various evaluations to explore the possible reasons and indications which impact the workload reshaping and prediction. The experimental results indicate that the principles of reshaping workload to improve prediction quality are constructing or improving the self-similarity of the reshaped workloads.

Magnetic susceptibility as a proxy for rainfall: Worldwide data from tropical and temperate climate

The magnetic susceptibility of soils has been linked with climate, mainly through rainfall, by numerous investigators. For this study we assembled a worldwide dataset of 272 samples, both published and new, to determine the nature of the relationship between MS and rainfall. On a worldwide basis there is a crude relationship ( r 2 = 0.265) between MS and rainfall. MS increases with increasing rainfall from about 200 mm/yr to 1000–1200 mm/yr. Above 1200 mm/yr, MS decreases as rainfall increases up to about 2000 mm/yr. Under arid and semi-arid conditions, below about 200 mm/yr of rainfall, MS and rainfall exhibit no relationship, likely the result of limited pedogenic activity. When tropical and temperate localities are analyzed separately, the correlation between MS and rainfall exhibits a dramatic improvement, r 2 = 0.568 and 0.520, respectively. For similar amounts of rainfall tropical localities generally exhibit lower MS values which may be related to differences in the seasonal variation of rainfall. Because the soil environment is highly variable, it is unlikely that a global or even regional equation could accurately estimate rainfall from MS. Using data from Morocco and Mali we demonstrate that local equations using only MS to estimate rainfall produce reasonable results, r 2 = 0.700 for Morocco and 0.611 for Mali. However, a single independent variable is unlikely to capture the variability of the complex soil system. To the local equations described above we experimented with adding several independent variables in addition to MS including diffuse reflectance spectral data (DRS), x-ray diffraction (XRD), and x-ray fluorescence (XRF). For DRS data, % reflectance and red, yellow, and near-infrared wavelengths appear important; for XRD illite and kaolinite counts are important whereas for XRF Fe 2O 3 is important. The addition of these other independent variables dramatically improves prediction quality. In the case of the Mali transect, estimating rainfall from these variables produces an r 2 that exceeds 0.9. Based on these data we conclude that MS has the potential to be a reasonable proxy of rainfall for mean annual precipitation values from 200 to 2000 mm/yr, especially if local equations are used. Predictions can be improved by including other independent variables that show sensitivity to climate and are related to MS through the chemistry of the weathering process; we included DRS, XRD, and XRF data with good results.

Calpain cleavage prediction using multiple kernel learning.

Calpain, an intracellular -dependent cysteine protease, is known to play a role in a wide range of metabolic pathways through limited proteolysis of its substrates. However, only a limited number of these substrates are currently known, with the exact mechanism of substrate recognition and cleavage by calpain still largely unknown. While previous research has successfully applied standard machine-learning algorithms to accurately predict substrate cleavage by other similar types of proteases, their approach does not extend well to calpain, possibly due to its particular mode of proteolytic action and limited amount of experimental data. Through the use of Multiple Kernel Learning, a recent extension to the classic Support Vector Machine framework, we were able to train complex models based on rich, heterogeneous feature sets, leading to significantly improved prediction quality (6% over highest AUC score produced by state-of-the-art methods). In addition to producing a stronger machine-learning model for the prediction of calpain cleavage, we were able to highlight the importance and role of each feature of substrate sequences in defining specificity: primary sequence, secondary structure and solvent accessibility. Most notably, we showed there existed significant specificity differences across calpain sub-types, despite previous assumption to the contrary. Prediction accuracy was further successfully validated using, as an unbiased test set, mutated sequences of calpastatin (endogenous inhibitor of calpain) modified to no longer block calpain's proteolytic action. An online implementation of our prediction tool is available at http://calpain.org.

Computational refinement of post-translational modifications predicted from tandem mass spectrometry

Motivation: A post-translational modification (PTM) is a chemical modification of a protein that occurs naturally. Many of these modifications, such as phosphorylation, are known to play pivotal roles in the regulation of protein function. Henceforth, PTM perturbations have been linked to diverse diseases like Parkinson's, Alzheimer's, diabetes and cancer. To discover PTMs on a genome-wide scale, there is a recent surge of interest in analyzing tandem mass spectrometry data, and several unrestrictive (so-called ‘blind’) PTM search methods have been reported. However, these approaches are subject to noise in mass measurements and in the predicted modification site (amino acid position) within peptides, which can result in false PTM assignments.Results: To address these issues, we devised a machine learning algorithm, PTMClust, that can be applied to the output of blind PTM search methods to improve prediction quality, by suppressing noise in the data and clustering peptides with the same underlying modification to form PTM groups. We show that our technique outperforms two standard clustering algorithms on a simulated dataset. Additionally, we show that our algorithm significantly improves sensitivity and specificity when applied to the output of three different blind PTM search engines, SIMS, InsPecT and MODmap. Additionally, PTMClust markedly outperforms another PTM refinement algorithm, PTMFinder. We demonstrate that our technique is able to reduce false PTM assignments, improve overall detection coverage and facilitate novel PTM discovery, including terminus modifications. We applied our technique to a large-scale yeast MS/MS proteome profiling dataset and found numerous known and novel PTMs. Accurately identifying modifications in protein sequences is a critical first step for PTM profiling, and thus our approach may benefit routine proteomic analysis.Availability: Our algorithm is implemented in Matlab and is freely available for academic use. The software is available online from http://genes.toronto.edu.Supplementary Information: Supplementary data are available at Bioinformatics online.Contact: frey@psi.utoronto.ca

Interval Type-2 TSK 퍼지논리시스템 기반 다중 퍼지 예측시스템 설계

본 논문은 예측 시스템의 성능을 개선하기 위해 비선형데이터의 내재된 특성이나 불확실성을 보다 효과적으로 반영할 수 있는 Interval Type-2 TSK 퍼지논리 시스템 기반 다중 퍼지 예측시스템의 설계를 다룬다. 본 논문에 제시된 다중 예측시스템들은 데이터의 비선형적 특성들을 효과적으로 고려하기 위해 설계되며, 각각의 시스템은 Type-1 TSK 퍼지논리나 다른 방법들에 비해 데이터의 불확실성을 충분히 반영할 수 있는 Interval Type-2 TSK 퍼지논리를 기반으로 구현된다. 또한, 1차 차분변환 과정을 통해, 데이터의 원형으로부터 최적의 차분데이터를 생성하고, 이들을 각 시스템의 입력으로 사용함으로써 시스템 설계 시 보다 안정된 통계적 정보를 제공할 수 있도록 한다. 마지막으로, 두 개의 전형적인 시계열 데이터의 예측 시뮬레이션을 통해 제안된 방법의 효용성을 검증한다. This paper presents multiple fuzzy prediction systems based on an Interval type-2 TSK fuzzy Logic System so that the uncertainty and the hidden characteristics of nonlinear data can be reflected more effectively to improve prediction quality. In proposed method, multiple fuzzy systems are adopted to handle the nonlinear characteristics of data, and each of multiple system is constructed by using interval type-2 TSK fuzzy logic because it can deal with the uncertainty and the characteristics of data better than type-1 TSK fuzzy logic and other methods. For input of each system, the first-order difference transformation method are used because the difference data generated from it can provide more stable statistical information to each system than the original data. Finally, computer simulations are performed to show the effectiveness of the proposed method for two typical time series examples.

Relabeling algorithm for retrieval of noisy instances and improving prediction quality

A relabeling algorithm for retrieval of noisy instances with binary outcomes is presented. The relabeling algorithm iteratively retrieves, selects, and re-labels data instances (i.e., transforms a decision space) to improve prediction quality. It emphasizes knowledge generalization and confidence rather than classification accuracy. A confidence index incorporating classification accuracy, prediction error, impurities in the relabeled dataset, and cluster purities was designed. The proposed approach is illustrated with a binary outcome dataset and was successfully tested on the standard benchmark four UCI repository dataset as well as bladder cancer immunotherapy data. A subset of the most stable instances (i.e., 7% to 51% of the sample) with high confidence (i.e., between 64%–99.44%) was identified for each application along with most noisy instances. The domain experts and the extracted knowledge validated the relabeled instances and corresponding confidence indexes. The relabeling algorithm with some modifications can be applied to other medical, industrial, and service domains.

Design of HCBKA-Based TSK Fuzzy Prediction System with Error Compensation

To improve prediction quality of a nonlinear prediction system, the system's capability for uncertainty of nonlinear data should be satisfactory. This paper presents a TSK fuzzy prediction system that can consider and deal with the uncertainty of nonlinear data sufficiently. In the design procedures of the proposed system, HCBKA(Hierarchical Correlationship-Based K-means clustering Algorithm) was used to generate the accurate fuzzy rule base that can control output according to input efficiently, and the first-order difference method was applied to reflect various characteristics of the nonlinear data. Also, multiple prediction systems were designed to analyze the prediction tendencies of each difference data generated by the difference method. In addition, to enhance the prediction quality of the proposed system, an error compensation method was proposed and it compensated the prediction error of the systems suitably. Finally, the prediction performance of the proposed system was verified by simulating two typical time series examples.

Multiparameter techniques for non-invasive measurement of blood glucose

Although painless glycemic control could improve life quality of patients, enabling a better regulation of hyper- and hypoglycaemia episodes, none of the available commercial devices offer enough precision to replace lancet approaches. The aim of this work is to describe the technologies of absorption spectroscopy and bioimpedance applied simultaneously to transcutaneous glycemic measurement. Basic electrical and optic theories were studied systematically and in following experiments glucose characteristics were measured in blood. Light spectroscopy studies were done in MIR ranges (5–40 μm), while complex bioimpedance measurements were scanned from 100 Hz to 30 MHz. Parallel to above parameters, information of epidermis temperature (conductance and IR radiation) was also collected, improving prediction quality. Cutaneous assays showed correlation of 0.43 for impedance and 0.46 for optic spectroscopy alone. The multiparameter analysis allowed the improvement of correlation to 0.57, after data treatment with Partial Least Square and neural networks method in Matlab.

Sequence based prediction of relative solvent accessibility using two-stage support vector regression with confidence values

Predicted relative solvent accessibility (RSA) provides useful information for prediction of binding sites and reconstruction of the 3D-structure based on a protein sequence. Recent years observed development of several RSA prediction methods including those that generate real values and those that predict discrete states (buried vs. exposed). We propose a novel method for real value prediction that aims at minimizing the prediction error when compared with six existing methods. The proposed method is based on a two-stage Support Vector Regression (SVR) predictor. The improved prediction quality is a result of the developed composite sequence representation, which includes a custom-selected subset of features from the PSI-BLAST profile, secondary structure predicted with PSI-PRED, and binary code that indicates position of a given residue with respect to sequence termini. Cross validation tests on a benchmark dataset show that our method achieves 14.3 mean absolute error and 0.68 correlation. We also propose a confidence value that is associated with each predicted RSA values. The confidence is computed based on the difference in predictions from the two-stage SVR and a second two-stage Linear Regression (LR) predictor. The confidence values can be used to indicate the quality of the output RSA predictions.

Support Vector Machines Approach to Predict the S&P CNX NIFTY Index Returns

The present study, investigates the predictability of S&P CNX NIFTY Index returns using Support vector machines (SVM). The performance of the SVM model in forecasting Nifty index returns is rigorously evaluated in terms of widely used statistical metrics like mean absolute error, root mean square error, normalized mean square error, correctness of sign and direction change (Pesaran and Timmermann (1992, DA test), and equal forecast accuracy using Diebold and Mariano (1995, DM test) by comparing its performance with those of neural network, random forest regression and a linear ARIMA model. The four competiting models are also examined in terms of various trading performance and economic criteria like annualized return, Sharpe ratio, maximum drawdown, annualized volatility, average gain/loss ratio etc via a trading experiment. The findings of the study reveal that SVM model achieves greater forecasting accuracy and improves prediction quality compared to other models experimented in the study. The SVM model can be used as an alternative forecasting tool for Nifty Index returns and it will lead to better returns based on the traditional forecasting accuracy measures, such as root mean squared errors, and financial criteria, such as risk-adjusted measures of return.

Feature based image processing methods applied to bathymetric measurements from airborne remote sensing in fluvial environments

AbstractBathymetric maps produced from remotely sensed imagery are increasingly common. However, when this method is applied to fluvial environments, changing scenes and illumination variations severely hinder the application of well established empirical calibration methods used to obtain predictive depth–colour relationships. In this paper, illumination variations are corrected with feature based image processing, which is used to identify areas in an image with a near‐zero water depth. This information can then be included in the depth–colour calibration process, which results in an improved prediction quality. The end product is an automated bathymetric mapping method capable of a 4 m2 spatial resolution with a precision of ±15 cm, which allows for a more widespread application of bathymetric mapping. Copyright © 2006 John Wiley & Sons, Ltd.

A collaborative filtering framework based on fuzzy association rules and multiple-level similarity

The rapid development of Internet technologies in recent decades has imposed a heavy information burden on users. This has led to the popularity of recommender systems, which provide advice to users about items they may like to examine. Collaborative Filtering (CF) is the most promising technique in recommender systems, providing personalized recommendations to users based on their previously expressed preferences and those of other similar users. This paper introduces a CF framework based on Fuzzy Association Rules and Multiple-level Similarity (FARAMS). FARAMS extended existing techniques by using fuzzy association rule mining, and takes advantage of product similarities in taxonomies to address data sparseness and nontransitive associations. Experimental results show that FARAMS improves prediction quality, as compared to similar approaches.