Determining Attribute Weights in a CBR Model for Early Cost Prediction of Structural Systems

This paper compares the performance of three optimization techniques, namely feature counting, gradient descent, and genetic algorithms (GA) in generating attribute weights that were used in a spreadsheet-based case based reasoning (CBR) prediction model. The generation of the attribute weights by using the three optimization techniques and the development of the procedure used in the CBR model are described in this paper in detail. The model was tested by using data pertaining to the early design parameters and unit cost of the structural system of 29 residential building projects. The results indicated that GA-augmented CBR performed better than CBR used in association with the other two optimization techniques. The study is of benefit primarily to researchers as it compares the impact attribute weights generated by three different optimization techniques on the performance of a CBR prediction tool.     

Wrapper Methods for Inductive Learning: Example Application to Bridge Decks

The decision tree algorithm is one of the most common techniques of inductive learning. This paper investigates the use of wrapper methods for bagging, boosting, and feature selection to improve the prediction accuracy of the decision tree algorithm. A set of concrete bridge decks is extracted from the Kansas bridge database, and the deterioration of the health index is selected as the decision/class value for induction. From the conducted experiments, the decision tree accuracy obtained is 67.7%, whereas bagging and the boosting gave 73.4% and 72.7%, respectively. Wrapping with a feature selection method gave an accuracy of 75.0%. If feature selection method is applied first, bagging and boosting do not provide any further improvement to the decision tree algorithm. A series of tests were conducted where the selected features were examined and manually eliminated for the data set. This revealed that the improvement obtained by the feature selection method can be misleading. For the problem at hand, the attributes selected were not the most important ones to the problem domain. Therefore, what may be an improvement from the machine learning or data mining viewpoint, can turn out to be a mistake from an engineering perspective. Automatically selected attributes should be checked carefully. Feature selection is not recommended in this case.     

Construction IT Decision Making Using Multiattribute Utility Theory for Use in a Laboratory Information Management System

This paper describes an objective process to select appropriate information technologies (ITs) for construction applications using the multiatttribute utility theory. Construction industry IT managers often find it difficult to objectively evaluate IT systems because each system has its own technical, economic, and risk considerations. This paper uses a multi-attribute utility theory to facilitate the decision-making process in such complex decisions. The theory’s basic premise is that the selection issue can be broken down into alternative attributes. Based upon the user’s tradeoffs among attributes, important weights are quantified and single-attribute utilities are measured. Finally, single-attribute utilities are combined to develop one single aggregate utility index for each alternative. A test is performed to use the utility theory to evaluate two different labeling approaches (radio-frequency identification and bar codes) for a construction material testing laboratory. Results showed that this technique has the ability to make decisions more objectively. Overall, the suggested model can also be adapted to evaluate other construction-related decisions, pertaining to other IT applications, equipment selection, and construction methods and techniques.     

CBR Revision Model for Improving Cost Prediction Accuracy in Multifamily Housing Projects

For any construction project to succeed, it is very important to accurately estimate the construction cost during the project’s initial stage. This is why there has been much interest lately in cost prediction models that use case-based reasoning (CBR). It has been pointed out, however, that existing CBR-based cost prediction models may yield inaccurate results even though they could survey optimal similar cases, if the number of cases in the case base is not enough. As opposed to the existing CBR-based construction cost prediction models, this study developed a CBR revision model that reflects the “revise” phase of the CBR cycle (retrieve, reuse, revise, and retain) based on nine multifamily housing projects executed recently by “A” Housing Corporation. To verify the developed model, a case study was performed using three case projects completed by “B” and “C” Housing Corporations. The result showed that the prediction error ratio after the Revise (I) phase decreased from 97.44 to 22.58%. This model can be effective when there are insufficient established cases in the case base.     

Uncertainty of Predictions of Embankment Dam Breach Parameters

Risk assessment studies considering the failure of embankment dams often require the prediction of basic geometric and temporal parameters of a breach, or the estimation of peak breach outflows. Many of the relations most commonly used to make these predictions were developed from statistical analyses of data collected from historic dam failures. The prediction uncertainties of these methods are widely recognized to be very large, but have never been specifically quantified. This paper presents an analysis of the uncertainty of many of these breach parameter and peak flow prediction methods. Application of the methods and the uncertainty analysis are illustrated through a case study of a risk assessment recently performed by the Bureau of Reclamation for a large embankment dam in North Dakota.     

基于正态隶属度—属性区间识别模型的岩爆倾向等级预测

岩爆是地下工程开挖面临的关键问题之一,为了准确预测深埋隧洞中岩爆烈度倾向等级,提出了正态隶属度—属性区间识别模型的岩爆预测方法。针对岩爆倾向等级属于典型的多属性有序分割类问题,构建了属性区间识别模型,并将岩爆倾向等级划分为4个等级进行预测。根据岩爆发生的成因和机理,选取应力系数、脆性系数、弹性应变指数和岩石完整性系数作为预测指标,考虑各指标之间、指标与标准等级之间的交互关系,采用正态隶属度函数和Jousselme距离计算评价指标权重。结合13个深埋隧洞工程对该预测模型进行准确性测试,并以双江口水电站SPD9厂房为例进行工程实例验证,该模型预测结果与实际相吻合,证明该模型用于具体工程实践中是可行且有效的,研究结果可为类似深埋隧洞岩爆倾向等级预测提供新的思路。     

Hybrid Neural Network Model for RH Vacuum Refining Process Control

A hybrid neural network model, in which RH process (theoretical) model is combined organically with neural network (NN) and case-base reasoning (CBR), was established.The CBR method was used to select the operation mode and the RH operational guide parameters for different steel grades according to the initial conditions of molten steel, and a three-layer BP neural network was adopted to deal with nonlinear factors for improving and compensating the limitations of technological model for RH process control and end-point prediction. It was verified that the hybrid neural network is effective for improving the precision and calculation efficiency of the model.     

Trade-offs depend on attribute range.

Contextual effects due to attribute range were examined in single-attribute and multiattribute judgments. The effect of a given attribute difference was greater when presented in a narrow range than a wide range. Stretching and shrinking the ranges of attributes altered the rank orders of judgments assigned to the same stimuli in different ranges. Trade-offs between time and money, between 1 measure of ability and another, and between achievement and motivation depended on attribute range. Although changes in trade-offs can result from changes in weights, data were consistent with the hypothesis that attribute range influences scale values. Scale values for common levels of an attribute spanned a wider interval when the attribute range was narrow than when the attribute range was wide. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Entropy Application to Improve Construction Finance Decisions

In financial decision-making processes, the adopted weights of the objective functions have significant impacts on the final decision outcome. However, conventional rating and weighting methods exhibit difficulty in deriving appropriate weights for complex decision-making problems with imprecise information. Entropy is a quantitative measure of uncertainty and has been useful in exploring weights of attributes in decision making. A fuzzy and entropy-based mathematical approach is employed to solve the weighting problem of the objective functions in an overall cash-flow model. The multiproject being undertaken by a medium-size construction firm in Hong Kong was used as a real case study to demonstrate the application of entropy. Its application in multiproject cash flow situations is demonstrated. The results indicate that the overall before-tax profit was HK$ 0.11 millions lower after the introduction of appropriate weights. In addition, the best time to invest in new projects arising from positive cash flow was identified to be two working months earlier than the nonweight system.     

Model for Predicting Performance of Architects and Engineers

Certain attributes of an architect or engineer (AE) may be used to predict his performance. These attributes may be categorized as “hard” or “soft” attributes. Hard attributes include an AE’s cognitive ability, job knowledge, task proficiency, and job experience. Soft attributes include an AE’s conscientiousness, initiative, social skills, controllability, and commitment. The purpose of this study is to identify those attributes that affect an AE’s performance, and to construct a model to predict his performance in design build (DB) projects. Twenty five attributes were generated using the hierarchy tree. The importance of these attributes was tested with designer/ builders who select and hire AEs, using a standard questionnaire. A statistical test showed that 24 of these attributes are significantly important. Thirty AEs were evaluated by experienced designer/builders (experts) who have worked with them in completed DB projects. Besides giving a global performance score of the AE (dependent variable), each expert also evaluated the AE on the degree to which they exhibited the important attributes (independent variables). Based on these ratings, an optimum multiple regression performance prediction model was obtained. To validate the model, another group of experts used the optimum model to evaluate 18 other AEs. The resulting performance score as calculated by the model was compared to the global performance scores awarded by the designer/builders. This validation process showed the model to be robust. The results of the study reveal that an AE’s performance can be predicted by using just three attributes: AE’s problem solving ability and project approach, AE’s speed in producing design drawings, and the AE’s level of enthusiasm in tackling a difficult assignment.     

End Temperature Prediction of Molten Steel in LF Based on CBR

In order to improve the temperature control level of molten steel in ladle furnace (LF), a case‐based reasoning (CBR) method has been proposed for predicting end temperature of molten steel in LF. To predict the temperature accurately and efficiently, this paper develops two‐step retrieval approach and the correlation based feature weighting (CFW) method for CBR. And, the study evaluates the prediction effect of CBR method by the experiment of comparison with back propagation neural network (BPNN) model and CBR model. Experimental results show that CBR model achieves better accuracy than BPNN model and the CBR method is effective to predict end temperature of molten steel in LF.     

通过化学分析预测锌基涂层耐蚀性能的专家系统

本研究旨在开发一种用于预测锌基涂层耐腐蚀性的通用方法,其可以表示为加速盐雾试验中的总质量损失。本方法仅基于三个分析参数,即锌、铝和镁的总涂层质量。这种限制的原因是这三种参数可能通过在线分析获得。然后,预测的耐腐蚀性被包括在一个过程/质量控制系统。加速腐蚀试验在布雷斯特的Swerea KIMAB IC（腐蚀研究所）以及比利时的冶金研究中心（CRM）进行。试验按照雷诺ECC1试验D172028/--C（12周）以及CRM研发的加速循环腐蚀试验进行。根据总质量损失情况,原材料被分为四个耐腐蚀级别。所有腐蚀试验都清晰、充分地说明了元素镁和铝的正面影响。对于涂层中大多数这些元素来说,元素镁和铝的影响比单独元素锌的影响大很多。因此,引入了一个新的量,叫做"等量元素锌涂层重量"。此量与锌、铝和镁的涂层重量线性相关。使用专家系统开发了一种用于预测耐腐蚀性的模型,此模型基于回归分析和"决策树"算法。根据上述提及的三个分析参数（即锌、铝和镁的总涂层质量）,可以使用开发的模型准确对27种材料中的25种进行分类。总之,即使是在线状态,这种方法也有可能准确预测出腐蚀行为。出于材料研发的目的,还扩展了专家系统使其包括其他分析参数。     

Formulating Procurement Selection Criteria through Case-Based Reasoning Approach

An appropriate procurement system is a catalyst to the success of a construction project. In practice, the solutions and outcomes of previous procurement selection decisions could be extremely useful in supporting decision making. As a technique that captures and reuses experiential knowledge, case-based reasoning (CBR) has a high potential for modeling the procurement selection decision within a complex dynamic environment. This paper examines the suitability of CBR approaches for procurement selection. The process involved in procurement selection is examined first. A conceptual framework for case-based procurement selection is proposed. The structure of a prototype model on procurement selection criteria (PSC) formulation is presented in this paper. The model applies the CBR approach to procurement criteria selection irrespective of the variability in the characteristics of the client, project, and external environment.     

Preliminary Cost Estimation Model Using Case-Based Reasoning and Genetic Algorithms

This study proposes a preliminary cost estimation model using case-based reasoning (CBR) and genetic algorithm (GA). In measuring similarity and retrieving similar cases from a case base for minimum prediction error, it is a key process in determining the factors with the greatest weight among the attributes of cases in the case base. Previous approaches using experience, gradient search, fuzzy numbers, and analytic hierarchy process are limited in their provision of optimal solutions. This study therefore investigates a GA for weight generation and applies it to real project data. When compared to a conventional construction cost estimation model, the accuracy of the CBR- and GA-based construction cost estimation model was verified. It is expected that a more reliable construction cost estimation model could be designed in the early stages by using a weight estimation technique in the development of a construction cost estimation model.     

Case-Based Reasoning Approach to Construction Safety Hazard Identification: Adaptation and Utilization

Risk assessment, consisting of hazard identification and risk analysis, is an important process that can prevent costly incidents. However, due to operational pressures and lack of construction experience, risk assessments are frequently poorly conducted. In order to improve the quality of risk assessments in the construction industry, it is important to explore the use of artificial intelligence methods to ensure that the process is efficient and at the same time thorough. This paper describes the adaptation process of a case-based reasoning (CBR) approach for construction safety hazard identification. The CBR approach aims to utilize past knowledge in the form of past hazard identification and incident cases to improve the efficiency and quality of new hazard identification. The overall approach and retrieval mechanism are described in earlier papers. This paper is focused on the adaptation process for hazard identification. Using the proposed CBR approach, for a new work scenario (the input case), a most relevant hazard identification tree and a set of incident cases will be retrieved to facilitate hazard identification. However, not all information contained in these cases are relevant. Thus, less relevant information has to be pruned off and all the retrieved information has to be integrated into a hazard identification tree. The proposed adaptation is conducted in three steps: (1) pruning of the retrieved hazard identification tree; (2) pruning of the incident cases; and (3) insertion of incident cases into the hazard identification tree. The adaptation process is based on the calculation of similarity scores of indexes. A case study based on actual hazard identifications and incident cases is used to validate the feasibility of the proposed adaptation techniques.     

Application of the one-dimensional Ising model to the secondary structures in globular proteins. Prediction of beta-regions

An attempt has been made to predict the beta-regions in 16 globular proteins by applying the one-dimensional Ising model theory of Lifson & Roig (8). The parameters for the theory have been derived from the statistical data on globular proteins given by Chou & Fasman (5). Comparison of our results with the data available from the X-ray crystallographic studies indicates a prediction accuracy which is comparable to those of several other methods, especially in view of the limitations in our method for considering the other secondary structures. It is pointed out that not considering the long-range interactions in our and other methods based on short-range interactions would make these methods incomplete and incapable of being uniformly applicable to all proteins.     

基于承钢生产数据预测烧结矿FeO含量北大核心CSCD

为了克服烧结矿中FeO含量检验滞后的问题,基于烧结生产各个环节所积累的大量数据,采用XGBoost算法建立FeO含量预测模型,以指导生产工作人员及时调整配料方案和设备参数。首先对承钢3号烧结机sqlsever数据库中的相关数据进行提取和整合,然后结合特征工程对特征参数数据进行一系列可视化分析和处理,最后将XGBoost算法应用于预测烧结矿FeO含量的建模当中,并与决策树模型预测效果进行对比。结果表明XGBoost模型预测效果较好,预测后的损失值最小可达0.071876,实现了准确预测FeO含量的目的,为烧结矿FeO含量的预测提供了一种有效的预测方法。     

Using evolutionary trees in protein secondary structure prediction and other comparative sequence analyses

Previously proposed methods for protein secondary structure prediction from multiple sequence alignments do not efficiently extract the evolutionary information that these alignments contain. The predictions of these methods are less accurate than they could be, because of their failure to consider explicitly the phylogenetic tree that relates aligned protein sequences. As an alternative, we present a hidden Markov model approach to secondary structure prediction that more fully uses the evolutionary information contained in protein sequence alignments. A representative example is presented, and three experiments are performed that illustrate how the appropriate representation of evolutionary relatedness can improve inferences. We explain why similar improvement can be expected in other secondary structure prediction methods and indeed any comparative sequence analysis method.     

The limitations of decision trees and automatic learning in real world medical decision making

The decision tree approach is one of the most common approaches in automatic learning and decision making. The automatic learning of decision trees and their use usually show very good results in various "theoretical" environments. But in real life it is often impossible to find the desired number of representative training objects for various reasons. The lack of possibilities to measure attribute values, high cost and complexity of such measurements, and unavailability of all attributes at the same time are the typical representatives. For this reason we decided to use the decision trees not for their primary task--the decision making--but for outlining the most important attributes. This was possible by using a well-known property of the decision trees--their knowledge representation, which can be easily understood by humans. In a delicate field of medical decision making, we cannot allow ourselves to make any inaccurate decisions and the "tips," provided by the decision trees, can be of a great assistance. Our main interest was to discover a predisposition to two forms of acidosis: the metabolic acidosis and respiratory acidosis, which can both have serious effects on child's health. We decided to construct different decision trees from a set of training objects. Instead of using a test set for evaluation of a decision tree, we asked medical experts to take a closer look at the generated trees. They examined and evaluated the decision trees branch by branch. Their comments show that trees generated from the available training set mainly have surprisingly good branches, but on the other hand, for some, no medical explanation could be found.