Centre and Range method for fitting a linear regression model to symbolic interval data

This paper introduces a new approach to fitting a linear regression model to symbolic interval data. Each example of the learning set is described by a feature vector, for which each feature value is an interval. The new method fits a linear regression model on the mid-points and ranges of the interval values assumed by the variables in the learning set. The prediction of the lower and upper bounds of the interval value of the dependent variable is accomplished from its mid-point and range, which are estimated from the fitted linear regression model applied to the mid-point and range of each interval value of the independent variables. The assessment of the proposed prediction method is based on the estimation of the average behaviour of both the root mean square error and the square of the correlation coefficient in the framework of a Monte Carlo experiment. Finally, the approaches presented in this paper are applied to a real data set and their performance is compared.     

Interval type-2 fuzzy expert system for prediction of carbon monoxide concentration in mega-cities

This paper presents an indirect approach to interval type-2 fuzzy logic system modeling to forecaste the level of air pollutants. The type-2 fuzzy logic system permits us to model the uncertainties among rules and the parameters related to data analysis. In this paper, we propose an indirect method to create an interval type-2 fuzzy logic system from a historical data, where Footprint of Uncertainties of fuzzy sets are extracted by implementation of an interval type-2 FCM algorithm and based on an upper and lower value for the level of fuzziness m in FCM. Finally, the proposed model is applied for prediction of carbon monoxide concentration in Tehran air pollution. It is shown that the proposed type-2 fuzzy logic system is superior in comparison to type-1 fuzzy logic systems in terms of two performance indices.     

A novel general approach to evaluating the PCBA for components with different membership function

The logistic support of high tech industry all emphasize that reliability design start, one must fully considers its life cycle. The reliability analysis and prediction are the main work objectives. The purpose is to ensure that the system (product) can achieve its designed functions under specific operating conditions. However, the incomplete failure data and different fault probability density function of the system elementary event also increase the difficulty of reliability design and calculation. It cannot be fully solved by traditional probability reliability. Therefore, a more general and efficient algorithm to assess the system reliability is needed. This paper proposed speculation by experts’ opinions according to incomplete information condition. It reasonably gives different fault membership function of possibility of failure distribution under different bottom event. It also applies fault-tree analysis, α-cut of vague set, and interval arithmetic operations of vague set to obtain fault interval and reliability interval of the system. Moreover, this paper also modifies Tanaka et al.’s fuzzy fault-tree definition and extended the new usage to fit different membership function of vague fault-tree. Then, find out the critical elementary event that affects the reliability of the system, which could be used for managerial decision-making, and as the bases to future system maintenance strategy. In numerical verification, a malfunction of printed circuit board assembly (PCBA) is presented as a numerical example. The result of the proposed method is compared with the listing approaches of reliability analysis methods.     

Deterministic and probabilistic ship pitch prediction using a multi-predictor integration model based on hybrid data preprocessing,reinforcement learning and improved QRNN

The deterministic and probabilistic prediction of ship motion is important for safe navigation and stable real-time operational control of ships at sea. However, the volatility and randomness of ship motion, the non-adaptive nature of single predictors and the poor coverage of quantile regression pose serious challenges to uncertainty prediction, making research in this field limited. In this paper, a multi-predictor integration model based on hybrid data preprocessing, reinforcement learning and improved quantile regression neural network (QRNN) is proposed to explore the deterministic and probabilistic prediction of ship pitch motion. To validate the performance of the proposed multi-predictor integrated prediction model, an experimental study is conducted with three sets of actual ship longitudinal motions during sea trials in the South China Sea. The experimental results indicate that the root mean square errors (RMSEs) of the proposed model of deterministic prediction are 0.0254°, 0.0359°, and 0.0188°, respectively. Taking series #2 as an example, the prediction interval coverage probabilities (PICPs) of the proposed model of probability predictions at 90%, 95%, and 99% confidence levels (CLs) are 0.9400, 0.9800, and 1.0000, respectively. This study signifies that the proposed model can provide trusted deterministic predictions and can effectively quantify the uncertainty of ship pitch motion, which has the potential to provide practical support for ship early warning systems.     

Prediction measurement with mean acceptable error for proper inconsistency in noisy weldability prediction data

Due to the complex nature of the welding process, the data used to construct prediction models often contain a significant amount of inconsistency. In general, this type of inconsistent data is treated as noise in the literature. However, for the weldability prediction, the inconsistency, which we describe as proper-inconsistency, may not be eliminated since the inconsistent data can help extract additional information about the process. This paper discusses that, in the presence of proper-inconsistency, it is inappropriate to perform the same approach generally employed with machine learning algorithms, in terms of the model construction and prediction measurement. Due to the numerical characteristics of proper-inconsistency, it is likely to achieve vague prediction results from the prediction model with the traditional prediction performance measures. In this paper, we propose a new prediction performance measure called mean acceptable error (MACE), which measures the performance of prediction models constructed with the presence of proper-inconsistency. This paper presents experimental results with real weldability prediction data, and we examine the prediction performance of k-nearest neighbor (kNN) and generalized regression neural network (GRNN) measured by MACE and the different characteristics of data in relation to MACE, kNN, and GRNN. The results indicate that using a smaller k on properly-inconsistent data increases the prediction performance measured by MACE. Also, the prediction performance on the correct data increases, while the effect of properly-inconsistent data decreases with the measurement of MACE.     

Probabilistic Boolean network modeling of an industrial machine

Theoretical modeling of manufacturing processes assists the design of new systems for predictions of future behavior, identifies improvement areas, and evaluates changes to existing systems. A novel approach is proposed to model industrial machines using probabilistic Boolean networks (PBNs) to study the relationship between machine components, their reliability and function. Once a machine is modeled as a PBN, through identification of regulatory nodes, predictors and selection probabilities, simulation and property verification are used to verify model correctness and behavior. Using real machine data, model parameters are estimated and a PBN is built to describe the machine, and formulate valid predictions about probability of failure through time. Two models were established: one with non-deterministic inputs (proposed), another with components’ MTBFs inputs. Simulations were used to generate data required to conduct inferential statistical tests to determine the level of correspondence between predictions and real machine data. An ANOVA test shows no difference between expected and observed values of the two models (p value = 0.208). A two-sample T test demonstrates the proposed model provides values closer to expected values; consequently, it can model observable phenomena (p value \(=\) 0.000). Simulations are used to generate data required to conduct inferential statistical tests to determine the level of correspondence between model prediction and real machine data. This research demonstrates that using PBNs to model manufacturing systems provides a new mechanism for the study and prediction of their future behavior at the design phase, assess future performance and identify areas to improve design reliability and system resilience.     

高速公路行程时间Bootstrap-KNN区间预测分析与实证

针对行程时间点预测不能描述预测结果的可信度问题,以高速公路收费系统作为基础数据源,提出基于Bootstrap的高速公路行程时间区间预测模型,通过范围概率(PICP)、预测区间平均宽度(MPIW)以及综合指标(CWC)反映区间预测性能.对预测模型建模和Bootstrap置信区间估计方法两个关键步骤进行分析和实证,比较小波神经网络和K最近邻两种常用数据驱动方法的预测误差,并分析4种Bootstrap置信区间估计方法的区间预测性能.在相同的置信水平下,Percentile Bootstrap-KNN模型的综合指标值CWC最小,说明该模型区间预测性能最佳.对陕西省高速公路某热点OD进行实例分析,结果表明,采用相同预测算法的区间预测比点预测的误差小,且预测区间宽度可以表征预测结果的可信度和参考价值.     

Identification of k-step-ahead prediction error model and MPC control

This work studies k-step-ahead prediction error model identification and its relationship to MPC control. The use of error criteria in parameter estimation will be discussed, where the identified model is used in model predictive control (MPC). Assume that the model error is dominated by the variance part, it can be shown that a k-step-ahead prediction error model is not optimal for k-step-ahead prediction. A normal one-step-ahead prediction error criterion will be optimal for k-step-ahead prediction. Then it is argued that even when some bias exists, the result could still hold true. Therefore, for MPC identification of linear processes, one-step-ahead prediction error models fever k-step-ahead prediction models. Simulations and industrial testing data will be used to illustrate the idea.     

RBF neural network inferential sensor for process emission monitoring

Inferential sensing, or soft sensing, gained popularity in recent years as an alternative to continuous emission monitoring systems because of its simplicity, reliability, and cost effectiveness as compared to analogous hardware sensors. In this paper we address the problem of NO_x emission using a model of furnace of an industrial boiler, and propose a neural network structure for high performance prediction of NO_x as well as O₂. The studied boiler is 160 MW, gas fired with natural gas, water-tube boiler, having two vertically aligned burners. The boiler model is a 3D problem that involves turbulence, combustion, radiation in addition to NO_x modeling. The 3D computational fluid dynamic model is developed using Fluent simulation package. The model provides calculations of the 3D temperature distribution as well as the rate of formation of the NO_x pollutant, enabling a better understanding on how and where NO_x are produced. The boiler was simulated under various operating conditions. The generated data is then used for initial development and assessment of neural network soft sensors for emission prediction based on the conventional process variable measurements. The performance of the proposed soft sensor is then evaluated using actual data from an industrial boiler. The developed soft sensor achieves comparable accuracy to the continuous emission monitor analyzer, however, with substantial reduction in the cost of equipment and maintenance.     

Using OVA modeling to improve classification performance for large datasets

One-Versus-All (OVA) classification is a classifier construction method where a k-class prediction task is decomposed into k 2-class sub-problems. One base model is constructed for each sub-problem and the base models are then combined into one model. Aggregate model implementation is the process of constructing several base models which are then combined into a single model for prediction. In essence, OVA classification is a method of aggregate modeling. This paper reports studies that were conducted to establish whether OVA classification can provide predictive performance gains when large volumes of data are available for modeling as is commonly the case in data mining. It is demonstrated in this paper that firstly, OVA modeling can be used to increase the amount of training data while at the same time using base model training sets whose size is much smaller than the total amount of available training data. Secondly, OVA models created from large datasets provide a higher level of predictive performance compared to single k-class models. Thirdly, the use of boosted OVA base models can provide higher predictive performance compared to un-boosted OVA base models. Fourthly, when the combination algorithm for base model predictions is able to resolve tied predictions, the resulting aggregate models provide a higher level of predictive performance.     

Rigorous modeling of gypsum solubility in Na–Ca–Mg–Fe–Al–H–Cl–H2O system at elevated temperatures

Precipitation and scaling of calcium sulfate have been known as major problems facing process industries and oilfield operations. Most scale prediction models are based on aqueous thermodynamics and solubility behavior of salts in aqueous electrolyte solutions. There is yet a huge interest in developing reliable, simple, and accurate solubility prediction models. In this study, a comprehensive model based on least-squares support vector machine (LS-SVM) is presented, which is mainly devoted to calcium sulfate dihydrate (or gypsum) solubility in aqueous solutions of mixed electrolytes covering wide temperature ranges. In this respect, an aggregate of 880 experimental data were gathered from the open literature in order to construct and evaluate the reliability of presented model. Solubility values predicted by LS-SVM model are in well accordance with the observed values yielding a squared correlation coefficient (R ²) of 0.994. Sensitivity of the model for some important parameters is also checked to ascertain whether the learning process has succeeded. At the end, outlier diagnosis was performed using the method of leverage value statistics to find and eliminate the falsely recorded measurements from assembled dataset. Results obtained from this study indicate that LS-SVM model can successfully be applied in predicting accurate solubility of calcium sulfate dihydrate in Na–Ca–Mg–Fe–Al–H–Cl–H₂O system over temperatures ranging from 283.15 to 371.15 K.     

Nonlinear regression applied to interval-valued data

This paper introduces a nonlinear regression model to interval-valued data. The method extends the classical nonlinear regression model in order to manage interval-valued datasets. The parameter estimates of the nonlinear model considers some optimization algorithms aiming to identify which one presents the best accuracy and precision in the prediction task. A detailed prediction performance study comparing the proposed nonlinear method and other linear regression methods for interval variables is presented based on K-fold cross-validation scheme with synthetic interval-valued datasets generated on a Monte Carlo framework. Moreover, two suitable real interval-valued datasets are considered to illustrate the usefulness and the performance of the approaches presented in this paper. The results suggested that the use of the nonlinear method is suitable for real datasets, as well as in the Monte Carlo simulation study.     

Conditional ordinal random fields for structured ordinal-valued label prediction

Predicting labels of structured data such as sequences or images is a very important problem in statistical machine learning and data mining. The conditional random field (CRF) is perhaps one of the most successful approaches for structured label prediction via conditional probabilistic modeling. In such models, it is traditionally assumed that each label is a random variable from a nominal category set (e.g., class categories) where all categories are symmetric and unrelated from one another. In this paper we consider a different situation of ordinal-valued labels where each label category bears a particular meaning of preference or order. This setup fits many interesting problems/datasets for which one is interested in predicting labels that represent certain degrees of intensity or relevance. We propose a fairly intuitive and principled CRF-like model that can effectively deal with the ordinal-scale labels within an underlying correlation structure. Unlike standard log-linear CRFs, learning the proposed model incurs non-convex optimization. However, the new model can be learned accurately using efficient gradient search. We demonstrate the improved prediction performance achieved by the proposed model on several intriguing sequence/image label prediction tasks.     

Prediction method of autoregressive moving average models for uncertain time series

ABSTRACT

Time series analysis is based on the continuous regularity of the development of objective things to predict the next value depending on observed values. Based on time series analysis, we present autoregressive moving average models to predict the next future value for an uncertain time series. In this paper, imprecise observations and disturbance terms are regarded as uncertain variables and assume that the latter are satisfied uncertain normal distribution. The prediction models of uncertain time series are established combining the knowledge of autoregressive model and uncertainty theory. Therefore, the interval range of the next future value is predicted based on the reliability constraint. As an illustration to compare with the numerical examples of the existing prediction method, the innovations and effectiveness of the work are further demonstrated by the computational results.     

An integrative model with subject weight based on neural network learning for bankruptcy prediction

This study proposes an integration strategy regarding how to efficiently combine the currently-in-use statistical and artificial intelligence techniques. In particular, by combining multiple discriminant analysis, logistic regression, neural networks, and decision trees induction, we introduce an integrative model with subject weight based on neural network learning for bankruptcy prediction. The strength of the proposed model stems from differentiating the weights of the source methods for each subject in the testing data set. That is, the relative weights consist of N by I matrix, where N denotes the number of subjects and I denotes the number of the source methods. The experiments using a real world financial data indicate that the proposed model can marginally increase the prediction accuracy compared to the source methods. The integration strategy can be useful for a dichotomous classification problem like bankruptcy prediction since prediction can be improved by taking advantage of existing and newly emerging techniques in the future.     

Reliability estimation using a genetic algorithm-based artificial neural network: An application to a load-haul-dump machine

In this study, a neural network-based model for forecasting reliability was developed. A genetic algorithm was applied for selecting neural network parameters like learning rate (η) and momentum (μ). The input variables of the neural network model were selected by maximizing the mean entropy value. The developed model was validated by applying two benchmark data sets. A comparative study reveals that the proposed method performs better than existing methods on benchmark data sets. A case study was conducted on a load-haul-dump (LHD) machine operated at a coal mine in Alaska, USA. Past time-to-failure data for the LHD machine were collected, and cumulative time-to-failure was calculated for reliability modeling. The results demonstrate that the developed model performs well with high accuracy (R² = 0.94) in the failure prediction of a LHD machine.     

Prediction of industrial equipment Remaining Useful Life by fuzzy similarity and belief function theory

We develop a novel prognostic method for estimating the Remaining Useful Life (RUL) of industrial equipment and its uncertainty. The novelty of the work is the combined use of a fuzzy similarity method for the RUL prediction and of Belief Function Theory for uncertainty treatment. This latter allows estimating the uncertainty affecting the RUL predictions even in cases characterized by few available data, in which traditional uncertainty estimation methods tend to fail. From the practical point of view, the maintenance planner can define the maximum acceptable failure probability for the equipment of interest and is informed by the proposed prognostic method of the time at which this probability is exceeded, allowing the adoption of a predictive maintenance approach which takes into account RUL uncertainty. The method is applied to simulated data of creep growth in ferritic steel and to real data of filter clogging taken from a Boiling Water Reactor (BWR) condenser. The obtained results show the effectiveness of the proposed method for uncertainty treatment and its superiority to the Kernel Density Estimation (KDE) and the Mean-Variance Estimation (MVE) methods in terms of reliability and precision of the RUL prediction intervals.     

Two-stage intonation modeling using feedforward neural networks for syllable based text-to-speech synthesis

This paper proposes a two-stage feedforward neural network (FFNN) based approach for modeling fundamental frequency (F₀) values of a sequence of syllables. In this study, (i) linguistic constraints represented by positional, contextual and phonological features, (ii) production constraints represented by articulatory features and (iii) linguistic relevance tilt parameters are proposed for predicting intonation patterns. In the first stage, tilt parameters are predicted using linguistic and production constraints. In the second stage, F₀ values of the syllables are predicted using the tilt parameters predicted from the first stage, and basic linguistic and production constraints. The prediction performance of the neural network models is evaluated using objective measures such as average prediction error (μ), standard deviation (σ) and linear correlation coefficient (γ_X,Y). The prediction accuracy of the proposed two-stage FFNN model is compared with other statistical models such as Classification and Regression Tree (CART) and Linear Regression (LR) models. The prediction accuracy of the intonation models is also analyzed by conducting listening tests to evaluate the quality of synthesized speech obtained after incorporation of intonation models into the baseline system. From the evaluation, it is observed that prediction accuracy is better for two-stage FFNN models, compared to the other models.     

Predicting mobility events on personal devices

High-end mobile phones are quickly becoming versatile sensing platforms, capable of continuously capturing the dynamic context of their owners through various sensors. A change in this context is often caused by the fact that owners–and therefore the devices they carry–are moving from one place to another. In this paper, we model the sensed environment as a stream of events, and assume, given that people are creatures of habit, that time correlations exists between successive events. We propose a method for the prediction in time of the next occurrence of an event of interest, such as ‘arriving at a certain location’ or ‘meeting with another person’, with a focus on the prediction of network visibility events as observed through the wireless network interfaces of the device. Our approach is based on using other events in the stream as predictors for the event we are interested in, and, in the case of multiple predictors, applying different strategies for the selection of the best predictor. Using two real-world data sets, we found that including predictors of infrequently occurring events results in better predictions using the best selection strategy. Also, we found that cross-sensor (cross-interface) information in most cases improves the prediction performance.     

A novel approach to designing an adaptive committee applied to predicting company’s future performance

This article presents an approach to designing an adaptive, data dependent, committee of models applied to prediction of several financial attributes for assessing company’s future performance. Current liabilities/Current assets, Total liabilities/Total assets, Net income/Total assets, and Operating Income/Total liabilities are the attributes used in this paper. A self-organizing map (SOM) used for data mapping and analysis enables building committees, which are specific (committee size and aggregation weights) for each SOM node. The number of basic models aggregated into a committee and the aggregation weights depend on accuracy of basic models and their ability to generalize in the vicinity of the SOM node. A random forest is used a basic model in this study. The developed technique was tested on data concerning companies from ten sectors of the healthcare industry of the United States and compared with results obtained from averaging and weighted averaging committees. The proposed adaptivity of a committee size and aggregation weights led to a statistically significant increase in prediction accuracy if compared to other types of committees.