关键词： Building energy   Energy retrofit   Predictive modeling   Machine learning   Urban sustainability  

摘要： Building energy conservation measures (ECMs) can significantly lower greenhouse gas (GHG) emissions from urban areas;however, uncertainties regarding not only ECM eligibility, but also associated costs and energy savings have slowed adoption of ECMs. To encourage ECM implementation, local governments have implemented a range of policies designed to increase the available information on building energy use. Energy audit mandates, such as New York City (NYC)'s Local Law 87 (LL87), require energy consultants to analyze installed building systems and provide building stakeholders with cost effective ECM recommendations on a multi-year cycle. However, complete audits are costly and time consuming. To accelerate ECM implementation, policymakers are exploring ways to utilize available data to target ECM5 across a city's entire building stock. In this study, energy audit data for over 1100 buildings in NYC, submitted in compliance with LL87, are analyzed to identify opportunities for ECMs across building system categories (e.g. distribution system, domestic hot water, etc.). A machine learning classifier, specifically a user-facing falling rule list (FRL) classifier based on binary features derived from LL87 data, is developed here to predict ECM eligibility given a specific set of building characteristics. Overall, the trained FRL classifier performs well (ROC AUC 0.72-0.86) for predicting cooling system, distribution system, domestic hot water, fuel switching, lighting, and motors ECM opportunities, which represent a majority of the auditor-recommended ECM5 in the sample. Additionally, linear decision lists developed by the model allow building stakeholders to easily conduct streamlined audits of building systems and identify possible ECM opportunities by limiting input to the most relevant factors and prioritizing likely retrofit candidates. The implications of this work are significant in accelerating the adoption of building ECMs and catalyzing energy use and GHG emiss

关键词： Support vector machines  

摘要： Side weirs are useful and common devices in flow measurement issues. These weirs have been extensively used in hydraulic and environmental engineering applications. The characteristic of flow over side weirs are completely different from that over a weir normal to the approach channel. An accurate flow modeling over a side weir mainly depends on the proper estimation of discharge coefficient. In this study, the potential of two different machine learning methods, namely support vector machines combined with genetic algorithm (SVM–GA) and gene expression programming (GEP) were evaluated for predicting trapezoidal and rectangular sharp-crested side weirs discharge coefficient. Correlation coefficient (R), mean normalize error (MNE), and Nash–Sutcliffe index (NS) statistics are used for the evaluation of the model’s performances. The results showed that the SVM–GA model with R = 0.97, NS = 0.94, and MNE = 13% for trapezoidal side weir and R = 0.97, NS = 0.91, and MNE = 18% for rectangular side weir, in test period gives more accurate results than GEP. The results verify that the SVM–GA model can process used data series by better generalization ability and higher prediction accuracy. © 2016 Indian Society for Hydraulics.

关键词： Multiple sclerosis   stationary wavelet entropy   decision tree   k-nearest neighbors   support vector machine   machine learning  

摘要： In order to detect multiple sclerosis (MS) subjects from healthy controls (HCs) in magnetic resonance imaging, we developed a new system based on machine learning. The MS imaging data was downloaded from the eHealth laboratory at the University of Cyprus, and the HC imaging data was scanned in our local hospital with volunteers enrolled from community advertisement. Inter-scan normalization was employed to remove the gray-level difference. We adjust the misclassification costs to alleviate the effect of unbalanced class distribution to the classification performance. We utilized two-level stationary wavelet entropy (SWE) to extract features from brain images. Then, we compared three machine learning based classifiers: the decision tree, k-nearest neighbors (kNN), and support vector machine. The experimental results showed the kNN performed the best among all three classifiers. In addition, the proposed SWE+kNN approach is superior to four state-of-the-art approaches. Our proposed MS detection approach is effective.

关键词： Resting-state functional magnetic resonance imaging(rs-fMRI)   Alzheimer's disease(AD)   Mild cognitive impairment(MCI)   Graph theory   Machine learning   Support vector machine (SVM)  

摘要： The study of brain networks by resting-state functional magnetic resonance imaging (rs-fMRI) is a promising method for identifying patients with dementia from healthy controls (HC). Using graph theory, different aspects of the brain network can be efficiently characterized by calculating measures of integration and segregation. In this study, we combined a graph theoretical approach with advanced machine learning methods to study the brain network in 89 patients with mild cognitive impairment (MCI), 34 patients with Alzheimer's disease (AD), and 45 age-matched HC. The rs-fMRI connectivity matrix was constructed using a brain parcellation based on a 264 putative functional areas. Using the optimal features extracted from the graph measures, we were able to accurately classify three groups (i.e., HC, MCI, and AD) with accuracy of 88.4 %. We also investigated performance of our proposed method for a binary classification of a group (e.g., MCI) from two other groups (e.g., HC and AD). The classification accuracies for identifying HC from AD and MCI, AD from HC and MCI, and MCI from HC and AD, were 87.3, 97.5, and 72.0 %, respectively. In addition, results based on the parcellation of 264 regions were compared to that of the automated anatomical labeling atlas (AAL), consisted of 90 regions. The accuracy of classification of three groups using AAL was degraded to 83.2 %. Our results show that combining the graph measures with the machine learning approach, on the basis of the rs-fMRI connectivity analysis, may assist in diagnosis of AD and MCI.

关键词： Machine learning   Deep learning   Text classification   Annotation of clinical text   Motivational interviewing  

摘要： This study examines the effectiveness of state-of-the-art supervised machine learning methods in conjunction with different feature types for the task of automatic annotation of fragments of clinical text based on codebooks with a large number of categories. We used a collection of motivational interview transcripts consisting of 11,353 utterances, which were manually annotated by two human coders as the gold standard, and experimented with state-of-art classifiers, including Naive Bayes, J48 Decision Tree, Support Vector Machine (SVM), Random Forest (RF), AdaBoost, DiscLDA, Conditional Random Fields (CRF) and Convolutional Neural Network (CNN) in conjunction with lexical, contextual (label of the previous utterance) and semantic (distribution of words in the utterance across the Linguistic Inquiry and Word Count dictionaries) features. We found out that, when the number of classes is large, the performance of CNN and CRF is inferior to SVM. When only lexical features were used, interview transcripts were automatically annotated by SVM with the highest classification accuracy among all classifiers of 70.8%, 61% and 53.7% based on the codebooks consisting of 17, 20 and 41 codes, respectively. Using contextual and semantic features, as well as their combination, in addition to lexical ones, improved the accuracy of SVM for annotation of utterances in motivational interview transcripts with a codebook consisting of 17 classes to 71.5%, 74.2%, and 75.1%, respectively. Our results demonstrate the potential of using machine learning methods in conjunction with lexical, semantic and contextual features for automatic annotation of clinical interview transcripts with near-human accuracy. (C) 2016 Elsevier Inc. All rights reserved.

关键词： Machine Learning cost model   Manufacturing cost   Economic production function   Design-to-cost   Gradient boosted trees   Support vector regression  

摘要： This paper compares the performance of five statistical models on the estimation of manufacturing cost of jet engine components, during the early design phase and using real industrial data. The analysis shows that recent techniques such as Gradient Boosted Trees and Support Vector Regression are up to two times more efficient than the ones typically encountered in the literature (Multiple Linear Regression and Artificial Neural Networks). If goodness-of-fit and predictive accuracy remain crucial to assess the performance of a model, other criteria such as computational cost, easiness to train or interpretability should be considered when selecting a statistical method for estimating the manufacturing cost of mechanical parts. Ideally, cost estimators should rely on several statistical models concurrently, as their distinct characteristics yield complementary views on the drivers of manufacturing cost. Finally, some engineering insights revealed by the statistical analysis are presented. They include the ranking and quantification of the most important cost drivers, the approximation of the economic production function of component cost according to accumulated production volume and a different view on the traditional breakdown of manufacturing cost of some jet engine components. As a conclusion, Machine Learning appears to be an effective, affordable, accurate and scalable technique to cost mechanical parts in the early stage of the design process. (C) 2016 Published by Elsevier B.V.

关键词： Machine learning   Mortality prediction   Abdominal aortic aneurysm   Clinical decision support system   Data analysis  

摘要： An abdominal aortic aneurysm is an abnormal dilatation of the aortic vessel at abdominal level. This disease presents high rate of mortality. and complications causing a decrease in the quality of life and increasing the cost of treatment. To estimate the mortality risk of patients undergoing surgery is complex due to the variables associated. The use of clinical decision support systems based on machine learning could help medical staff to improve the results of surgery and get a better understanding of the disease. In this work, the authors present a predictive system of inhospital mortality in patients who were undergoing to open repair of abdominal aortic aneurysm. Different methods as multilayer perceptron, radial basis function and Bayesian networks are used. Results are measured in terms of accuracy, sensitivity and specificity of the classifiers, achieving an accuracy higher than 95%. The developing of a system based on the algorithms tested can be useful for medical staff in order to make a better planning of care and reducing undesirable surgery results and the cost of the post-surgical treatments. (C) 2016 Elsevier Inc. All rights reserved.

关键词： Algorithms   Applied mathematics   Artificial intelligence   Biology and life sciences   Cognitive science   Communications   Computer and information sciences   Earth sciences   Geographic information systems   Geography   Geoinformatics   Infectious diseases   Influenza   Machine learning   Machine learning algorithms   Mathematics   Medicine and health sciences   Network analysis   Neuroscience   Physical sciences   Public and occupational health   Research and analysis methods   Research Article   Simulation and modeling   Social communication   Social media   Social networks   Social sciences   Sociology   Support vector machines   Twitter   Viral diseases  

摘要： Traditional methods for monitoring influenza are haphazard and lack fine-grained details regarding the spatial and temporal dynamics of outbreaks. Twitter gives researchers and public health officials an opportunity to examine the spread of influenza in real-time and at multiple geographical scales. In this paper, we introduce an improved framework for monitoring influenza outbreaks using the social media platform Twitter. Relying upon techniques from geographic information science (GIS) and data mining, Twitter messages were collected, filtered, and analyzed for the thirty most populated cities in the United States during the 2013-2014 flu season. The results of this procedure are compared with national, regional, and local flu outbreak reports, revealing a statistically significant correlation between the two data sources. The main contribution of this paper is to introduce a comprehensive data mining process that enhances previous attempts to accurately identify tweets related to influenza. Additionally, geographical information systems allow us to target, filter, and normalize Twitter messages.

关键词： Oral mucositis   NTCP modelling   Dose-response modelling   Machine learning   Spatial dose metrics   Head and neck radiotherapy  

摘要： Background and purpose: Severe acute mucositis commonly results from head and neck (chemo)radiotherapy. A predictive model of mucositis could guide clinical decision-making and inform treatment planning. We aimed to generate such a model using spatial dose metrics and machine learning. Materials and methods: Predictive models of severe acute mucositis were generated using radiotherapy dose (dose-volume and spatial dose metrics) and clinical data. Penalised logistic regression, support vector classification and random forest classification (RFC) models were generated and compared. Internal validation was performed (with 100-iteration cross-validation), using multiple metrics, including area under the receiver operating characteristic curve (AUC) and calibration slope, to assess performance. Associations between covariates and severe mucositis were explored using the models. Results: The dose-volume-based models (standard) performed equally to those incorporating spatial information. Discrimination was similar between models, but the RFCstandard had the best calibration. The mean AUC and calibration slope for this model were 0.71 (s.d. = 0.09) and 3.9 (s.d. = 2.2), respectively. The volumes of oral cavity receiving intermediate and high doses were associated with severe mucositis. Conclusions: The RFCstandard model performance is modest-to-good, but should be improved, and requires external validation. Reducing the volumes of oral cavity receiving intermediate and high doses may reduce mucositis incidence. (C) 2016 The Author(s). Published by Elsevier Ireland Ltd.

关键词： Weather prediction   Solar irradiance   Energy efficiency   Sustainability   Resilience   Support vector machine regression   Probability estimation  

摘要： Accurate weather information is required to assess the performance of sustainable buildings. Currently such performance is assessed using Typical Meteorological Year (TMY) weather files. TMY3 is the latest TMY file introduced by Department of Energy (DOE) and it represents the historical 30-year average of weather data more closely than any other available datasets. However, TMY3 data reflects the past and do not accurately forecast weather information thus leading to erroneous estimation of performance and economic feasibility. Therefore, more accurate methods are necessary to generate weather files so as to design buildings for sustainability and resilience. This article proposes an effective hybrid modeling methodology based on support vector machine regression and probability estimation to predict long term future weather variables. The methodology has been validated by using a dataset containing historical solar irradiance data for 54 years (1961-2014) for Cincinnati, Ohio. By implementing the proposed technique multiple models were trained with different segments of the dataset, which were used to predict the hourly solar irradiance for 8, 9, 10 and 11 years respectively. Subsequently, the predicted data and the current TMY3 data was compared with the actual historical data. The average increase in accuracy of the predicted over TMY3 direct normal and diffuse horizontal components of solar irradiance was found to be 38.43% and 11.8%. These findings suggest that the long term accuracy of the proposed modeling technique is greater than the current state-of-the-art. (C) 2016 Elsevier Ltd. All rights reserved.