关键词： Magnetic resonance imaging   Prostatic neoplasms   Neoplasm grading   Radiomics  

摘要： Purpose: To evaluate the performance of a multi-parametric MRI (mp-MRI)-based radiomics signature for discriminating between clinically significant prostate cancer (csPCa) and insignificant PCa (ciPCa). Materials and methods: Two hundred and eighty patients with pathology-proven PCa were enrolled and were randomly divided into training and test cohorts. Eight hundred and nineteen radiomics features were extracted from mp-MRI for each patient. The minority group in the training cohort was balanced via the synthetic minority over-sampling technique (SMOTE) method. We used minimum-redundancy maximum-relevance (mRMR) selection and the LASSO algorithm for feature selection and radiomics signature building. The classification performance of the radiomics signature for csPCa and ciPCa was evaluated by receiver operating characteristic curve analysis in the training and test cohorts. Results: Nine features were selected for the radiomics signature building. Significant differences in the radiomics signature existed between the csPCa and ciPCa groups in both the training and test cohorts (p < 0.01 for both). The AUC, sensitivity and specificity of the radiomics signature were 0.872 (95% CI: 0.823-0.921), 0.883, and 0.753, respectively, in the training cohort, and 0.823 (95% CI: 0.669-0.976), 0.841, and 0.727, respectively, in the test cohort. Conclusion: Mp-MRI-based radiomics signature have the potential to noninvasively discriminate between csPCa and ciPCa.

关键词： Animal breeding   animal health   machine learning   prediction   regression  

摘要： The current livestock management landscape is transitioning to a high-throughput digital era where large amounts of information captured by systems of electro-optical, acoustical, mechanical, and biosensors is stored and analyzed on a daily and hourly basis, and actionable decisions are made based on quantitative and qualitative analytic results. While traditional animal breeding prediction methods have been used with great success until recently, the deluge of information starts to create a computational and storage bottleneck that could lead to negative long-term impacts on herd management strategies if not handled properly. A plethora of machine learning approaches, successfully used in various industrial and scientific applications, made their way in the mainstream approaches for livestock breeding techniques, and current results show that such methods have the potential to match or surpass the traditional approaches, while most of the time they are more scalable from a computational and storage perspective. This article provides a succinct view on what traditional and novel prediction methods are currently used in the livestock breeding field, how successful they are, and how the future of the field looks in the new digital agriculture era.

关键词： 机器学习   多因子选股   XGBoost   模型   参数寻优   IC   值法  

摘要： 近年来，机器学习方法在量化投资的应用研究逐渐兴起，但关于新兴的机器学习模型(XGBoost 模型) 在量化选股方面的研究并不算多，本论文以沪深300成分股作为样本对XGBoost 模型进行多因子选股研究，并结合支持向量机模型、随机森林模型、线性回归模型比较分析各模型在选股性能上的优劣。 首先，在第一章中指出传统多因子模型的有效性随着因子池的扩大而不断降低，而各种机器学习模型都是一种有效的解决方案，其中在理论上具有优越性的XGBoost 模型也有待实证进一步检验，为全论文奠定了理论基础，具有研究意义。 其次，在因子选择方面，本论文从价值投资理念出发并结合行为金融学的思想，选择了估值类、财务类、成长类、动量类及情绪类共5 大类20 个因子，较为全面的反映了影响股票收益率的因素。为进一步优化因子池进而为模型实证创造最佳样本数据，本论文还对选取的初始数据进行数据预处理及单因子有效性检验。本论文还综合了中位数去极值法、均值填充法、行业市值中性化处理及数据标准化等方法对原始数据进行预处理进而得到了标准数据格式；在单因子有效性检验中，本论文综合运用了IC 值法及分组收益法对单个因子的有效性进行综合评价，得到了11 个有效因子，然后再根据相关系数法对这11 个因子进行冗余因子剔除并最终得到了9 个有效因子。 然后，本论文采用网格寻优法及交叉验证法对XGBoost 模型进行参数寻优并得出最优参数。并且，根据各因子得分权重值可以看出其中momenturn_1m(个股最近1 个月收益率) 与Attention_rate (1 月关注增长率) 等反映市场情绪的短期因子重要性排名靠前，而诸如Sales_G (营业收入同比增长率) 等反映基本面的长期因子重要性排名靠后，在一定程度上也反映了我国股市短期因素主导较强、投机较为严重，中长线价值投资理念还没有很好被普及和贯彻执行。 最后，本论文对XGBoost 模型及各模型进行了回测检验，并对模型损失函数进行了改进，同时运用收益-风险评价体系比较各模型选股性能的优劣；此外本论文进行了三次回测以纵向观测时间维度对模型有效性的影响。回测结果表明：各模型在多因子选股上均能产生超额收益，这在一定程度上肯定了本文因子选取的有效性；从各模型的横向比较来看，XGBoost 模型整体表现优于其他模型，这从实证上验证了XGBoost 模型强大的集成学习能力，同时结果表明改进损失函数有效的提升了模型的预测能力；从三次回测结果的纵向比较来看，随着回测区间的增长，各模型的Alpha 收益显著降低，并且风险指标也在不断增大，表明模型预测的稳定性会随着时间的延长而减弱。

摘要： Clinical risk stratification for sudden cardiac death (SCD) in hypertrophic cardiomyopathy (HC) employs rules derived from American College of Cardiology Foundation/American Heart Association (ACCF/AHA) guidelines or the HCM Risk-SCD model (C-index similar to 0.69), which utilize a few clinical variables. We assessed whether data-driven machine learning methods that consider a wider range of variables can effectively identify HC patients with ventricular arrhythmias (VAr) that lead to SCD. We scanned the electronic health records of 711 HC patients for sustained ventricular tachycardia or ventricular fibrillation. Patients with ventricular tachycardia or ventricular fibrillation (n = 61) were tagged as VAr cases and the remaining (n = 650) as non-VAr. The 2-sample t test and information gain criterion were used to identify the most informative clinical variables that distinguish VAr from non-VAr;patient records were reduced to include only these variables. Data imbalance stemming from low number of VAr cases was addressed by applying a combination of over- and under sampling strategies. We trained and tested multiple classifiers under this sampling approach, showing effective classification. We evaluated 93 clinical variables, of which 22 proved predictive of VAr. The ensemble of logistic regression and naive Bayes classifiers, trained based on these 22 variables and corrected for data imbalance, was most effective in separating VAr from non-VAr cases (sensitivity = 0.73, specificity = 0.76, C-index = 0.83). Our method (HCM-VAr-Risk Model) identified 12 new predictors of VAr, in addition to 10 established SCD predictors. In conclusion, this is the first application of machine learning for identifying HC patients with VAr, using clinical attributes. Our model demonstrates good performance (C-index) compared with currently employed SCD prediction algorithms, while addressing imbalance inherent in clinical data. (C) 2019 The Authors. Published by Elsevier Inc.

关键词： Water quality   Spectrophotometry   Early contamination detection   Event detection system  

摘要： Water is a resource that affects every aspect of life. Intentional (terrorist or wartime events) or accidental water contamination events could have a tremendous impact on public health, behavior and morale. Quick detection of such events can mitigate their effects and reduce the potential damage. Continuous on-line monitoring is the first line of defense for reducing contamination associated damage. One of the available tools for such detection is the UV-absorbance spectrophotometry, where the absorbance spectra are compared against a set of normal and contaminated water fingerprints. However, as there are many factors at play that affect this comparison, it is an elusive and tedious task. Further, the comparison against a set of known fingerprints is futile when the water in the supply system are a mix, with varying proportions, of water from different sources, which differ significantly in their physicochemical characteristics. This study presents a new scheme for early detection of contamination events through UV absorbance under unknown routine conditions. The detection mechanism is based on a new affinity measure, Fitness, and a procedure similar to Gram based amplification, which result in a flexible mechanism to alert if a contamination is present. The method was shown to be most effective when applied to a set of comprehensive experiments, which examined the absorbance of various contaminants in drinking water in lab and real-life configurations. Four datasets, which contained real readings from either laboratory experiments or monitoring station of an operational water supply system were used. To extend the testbed even further, an artificial dataset, simulating a vast array of proportions between specific water sources is also presented. The results show, that for all datasets, high detection rates, while maintaining low levels of false alarms, were obtained by the algorithm. (C) 2019 Elsevier Ltd. All rights reserved.

关键词： Water reuse   Potable water   Cross-connection   Water quality parameter   Real-time detection   Machine learning method  

摘要： Risk of cross-connection is becoming higher due to greater construction of potable-reclaimed water dual distribution systems. Cross-connection events can result in serious health concerns and reduce public confidence in reclaimed water. Thus, reliable, cost-effective and real-time online detection methods for early warning are required. The current study carried out pilot-scale experiments to simulate potable-reclaimed water pipe cross-connection events for different mixing ratios (from 30% to 1%) using machine learning methods based on multiple conventional water quality parameters. The parameters included residual chlorine, pH, turbidity, temperature, conductivity, oxidation-reduction potential and chemical oxygen demand. The results showed that correlated variation occurred among water quality parameters at the time of the cross-connection event. A single parameter-based method can be effective at high mixing ratios, but not at low mixing ratios. The direct supporting vector machine (SVM)-based method managed to overcome this drawback, but coped poorly with abnormal readings of water parameter sensors. In that respect, a Pearson correlation coefficient (PCC)-SVM-based method was developed. It provided not only high detection performance under normal conditions, but also remained reliable when abnormal readings occurred. The detection accuracy and true positive rate of this method was still over 88%, and the false positive rate was below 12%, given a sudden variation of an individual water quality parameter. The receiver operating characteristic curves further confirmed the promising practical applicability of this PCC-SVM-based method for early detection of cross-connection events.

关键词： aerosol jet printing   direct writing   hybrid machine learning   line morphology   operating process window   quality optimization  

摘要： Aerosol jet printing (AJP) is a three-dimensional (3D) noncontact and direct printing technology for fabricating customized microelectronic devices on flexible substrates. Despite the capability of fine feature deposition, the complicated relationship between the main process parameters will affect the printing quality significantly in a design space. In this paper, a novel hybrid machine learning method is proposed to determine the optimal operating process window for the AJP process in various design spaces. The proposed method consists of classic machine learning methods, including experimental sampling, data clustering, classification, and knowledge transfer. In the proposed method, a two-dimensional design space is fully explored by a Latin hypercube sampling experimental design at a certain print speed. Then, the influence of the sheath gas flow rate (SHGFR) and the carrier gas flow rate (CGFR) on the printed line quality is analyzed by a K-means clustering approach, and an optimal operating process window is determined by a support vector machine. To efficiently identify more operating process windows at different print speeds, a transfer learning approach is applied to exploit relatedness between different operating process windows. Hence, at a new print speed, the number of line samples for identifying a new operating process window is greatly reduced. Finally, to balance the complex relationship among SHGFR, CGFR, and print speed, a 3D operating process window is determined by an incremental classification approach. Different from experiment-based approaches adopted in 3D printing technologies for quality optimization, the proposed method is developed based on the theory of knowledge discovery and data mining. Therefore, the knowledge in different design spaces can be fully explored and transferred for printed line quality optimization. Moreover, the data-driven-based characteristics can help the proposed method develop a guideline for quality optimization

关键词： augmentation   asymptotic efficiency   cross-validation   influence function   semiparametric theory   super learner  

摘要： Clinical trials are widely considered the gold standard for treatment evaluation, and they can be highly expensive in terms of time and money. The efficiency of clinical trials can be improved by incorporating information from baseline covariates that are related to clinical outcomes. This can be done by modifying an unadjusted treatment effect estimator with an augmentation term that involves a function of covariates. The optimal augmentation is well characterized in theory but must be estimated in practice. In this article, we investigate the use of machine learning methods to estimate the optimal augmentation. We consider and compare an indirect approach based on an estimated regression function and a direct approach that aims directly to minimize the asymptotic variance of the treatment effect estimator. Theoretical considerations and simulation results indicate that the direct approach is generally preferable over the indirect approach. The direct approach can be implemented using any existing prediction algorithm that can minimize a weighted sum of squared prediction errors. Many such prediction algorithms are available, and the super learning principle can be used to combine multiple algorithms into a super learner under the direct approach. The resulting direct super learner has a desirable oracle property, is easy to implement, and performs well in realistic settings. The proposed methodology is illustrated with real data from a stroke trial.

关键词： urbanization   urban extent   urban land use   remote sensing   machine learning   VIIRS   MODIS  

摘要： Urbanization poses significant challenges on sustainable development, disaster resilience, climate change mitigation, and environmental and resource management. Accurate urban extent datasets at large spatial scales are essential for researchers and policymakers to better understand urbanization dynamics and its socioeconomic drivers and impacts. While high-resolution urban extent data products - including the Global Human Settlements Layer (GHSL), the Global Man-Made Impervious Surface (GMIS), the Global Human Built-Up and Settlement Extent (HBASE), and the Global Urban Footprint (GUF) - have recently become available, intermediate-resolution urban extent data products including the 1 km SEDAC's Global Rural-Urban Mapping Project (GRUMP), MODIS 1km, and MODIS 500 m still have many users and have been demonstrated in a recent study to be more appropriate in urbanization process analysis (around 500 m resolution) than those at higher resolutions (30 m). The objective of this study is to improve large-scale urban extent mapping at an intermediate resolution (500 m) using machine learning methods through combining the complementary nighttime Visible Infrared Imaging Radiometer Suite (VIIRS) and daytime Moderate Resolution Imaging Spectroradiometer (MODIS) data, taking the conterminous United States (CONUS) as the study area. The effectiveness of commonly-used machine learning methods, including random forest (RF), gradient boosting machine (GBM), neural network (NN), and their ensemble (ESB), has been explored. Our results show that these machine learning methods can achieve similar high accuracies across all accuracy metrics (>95% overall accuracy, >98% producer's accuracy, and >92% user's accuracy) with Kappa coefficients greater than 0.90, which have not been achieved in the existing data products or by previous studies;the ESB is not able to produce significantly better accuracies than individual machine learning methods;the total misclassifications generated by GB

关键词： suburban land cover classification   LiDAR   aerial imagery   machine learning   nDSM   surface roughness   surface intensity  

摘要： It is valuable to study the land use/land cover (LULC) classification for suburbs. The fusion of Light Detection and Ranging (LiDAR) data and aerial imagery is often regarded as an effective method for the LULC classification;however, more in-depth analysis would be required to explore effective information for enhancing the suburban LULC classification. In this study, first, both aerial imageries and point clouds were simultaneously collected. Then, LiDAR-derived models, i.e., normalized digital surface model (nDSM) and surface intensity model (IM), were generated from the elevation and intensity of point clouds. Further, considering the surface characteristics of ground objects in suburb, we proposed a new LiDAR-derived model, namely surface roughness model (RM), to reveal the degree of surface fluctuations. Additionally, various combinations of aerial imageries and the LiDAR-derived data were used to analyze the effects of multi-variable fusion under different scenarios and optimize the multi-variable integration for suburban LULC classification. The mean decrease impurity method was used to identify the importance of variables;three machine learning classifiers, i.e., random forest (RF), k-nearest neighbor (KNN) and artificial neural network (ANN) were adopted in various scenarios. The results were as follows. The fusion of aerial imagery and all the LiDAR-derived models, i.e., nDSM, RM and IM, with RF classifier performs best in the suburban LULC classification (overall accuracy = 84.75%, kappa coefficient = 0.80). Variable importance analysis shows that nDSM has the highest variable importance proportion (VIP) value, followed by RM, IM, and spectral information, indicating the feasibility of this proposed LiDAR-derived model-RM. This research presents effective methods relating to the application of aerial imagery and LiDAR-derived model for the complex suburban surface scenarios.