关键词： OWOM   消费者评论评分   有效性评分   文本挖掘   机器学习  

摘要： 随着当今主要旅游网站的竞争愈演愈烈,争夺潜在消费者与维护已有用户已经成为进驻旅游网站的酒店平台的强烈需求。正因为良好的网络口碑(OWOM)传播会给酒店平台和旅游网站带来良好的声誉与可观的收益,由生产消费者(Prosumers)所留下的网络口碑评论生成(Review Generation)和评论消费(Review Consumption)的相关信息需要被密切关注。鉴于网络口碑的重要性,对于酒店管理者和旅游平台来说,判断消费者评论的哪些方面与其评论的满意度密切相关就变得至关重要。正因如此,我们聚焦于消费者评论文本内容所展现的多种维度信息。在本论文中,我们用到了机器学习方法,来探究是否能通过定性的网络口碑评论文本内容来正确地对一条评论的好坏程度做分类。特别地,我们挖掘了网络口碑文本内容中所隐含的情感极性与情感主观度等多种维度与观点的信息。通过用五种机器学习算法构建的二分类分类器,我们尝试对网络口碑定量的消费者评论评分和评论有效性评分做预测。我们的预测结果显示,文本中所隐含的情感极性和七种维度的观点特征对评论评分可以做出很好的预测,这对酒店管理者和旅游网站平台研究其中的微观细节并更好地相应客户需求有深刻的意义。同时,我们的特征没有很好地预测评论有效性评分,这说明在有效性评分的时间累积效应之外,目前从文本内容中所提取到的特征并不能对其产生有效影响,可能还有其他尚未发现的有效因素。

关键词： integrated optics   silicon photonics   nanophotonics   machine learning   artificial neural networks  

摘要： Interest in nanophotonics continues to grow as integrated optics provides an affordable platform for areas like telecommunications, quantum information processing, and biosensing. Designing and characterizing integrated photonics components and circuits, however, remains a major bottleneck. This is especially true when complex circuits or devices are required to study a particular phenomenon. To address this challenge, this work develops and experimentally validates a novel machine learning design framework for nanophotonic devices that is both practical and intuitive. As case studies, artificial neural networks are trained to model strip waveguides, integrated chirped Bragg gratings, and microring resonators using a small number of simple input and output parameters relevant to designers. Once trained, the models significantly decrease the computational cost relative to traditional design methodologies. To illustrate the power of the new design paradigm, both forward and inverse design tools enabled by the new design paradigm are demonstrated. These tools are directly used to design and fabricate several integrated Bragg grating devices and ring resonator filters. The method's predictions match the experimental measurements well and do not require any post-fabrication training adjustments.

关键词： Medical imaging   Artificial intelligence  

摘要： Energy-resolved photon counting detectors have potential improvements over conventional scintillating x-ray detectors and can be used for spectral computed tomography (CT) applications such as material decomposition. However, photon counting detector are difficult to model because they have many non-ideal physical phenomena. We developed machine learning algorithms to model the spectral measurements of energy-resolved photon counting detectors for the purposes of material decomposition and image reconstruction. A feed-forward neural network was used to estimate path lengths of different materials that the x-ray path traversed from spectral measurements of a photon counting detector using calibration measurements on an experimental bench-top CT system. The neural network relied only on calibration measurements and no prior knowledge of the source spectrum, material attenuation coefficients, or detector response. The neural networks decomposed each ray in the measured sinograms into material separated sinograms and reconstructed into basis material images. The capabilities of the machine learning algorithm were evaluated experimentally with a rod phantom with calculated biases of 0.3% to 7.6% compared to biases of 1.3% to 16% of a previously proposed empirical decomposition algorithm. A special case of material decomposition occurs when materials with high atomic numbers are in the measured ray path. These materials have K-absorption edges in their linear attenuation coefficients in the relevant x-ray energy range. In a separate study, the neural network mode was extended for K-edge imaging to separate contrast agents and nanoparticles, such as iodine and gadolinium, from other materials in the image. Transfer learning was used to further train the neural networks between detector pixels and from simulated calibration data to improve the qualitative and quantitative image quality in a rod phantom and an ex-vivo rat leg specimen. Similarly, a neural network was develop

关键词： Machine learning   data mining   classification   incomplete data   imbalanced data   Bayesian networks   acute myocardial infarction  

摘要： The heart is an important organ in the human body, and acute myocardial infarction (AMI) is the leading cause of death in most countries. Researchers are doing a lot of data analysis work to assist doctors in predicting the heart problem. An analysis of the data related to different health problems and its functions can help in predicting the wellness of this organ with a degree of certainty. Our research reported in this paper consists of two main parts. In the first part of the paper, we compare different predictive models of hospital mortality for patients with AMI. All results presented in this part are based on real data of about 603 patients from a hospital in the Czech Republic and about 184 patients from two hospitals in Syria. Although the learned models may be specific to the data, we also draw more general conclusions that we think are generally valid. In the second part of the paper, because the data is incomplete and imbalanced we develop the Chow-Liu and tree-augmented naive Bayesian to deal with that data in better conditions, and compare the quality of these algorithms with others.

关键词： Biology   Biostatistics   Genetics   Artificial intelligence  

摘要： For many diseases, especially cancer, patients suffer from a lack of response or acquired resistance to current approved therapeutics. Therefore, the need to rapidly identify novel therapies is pertinent, however time intensive experimental screenings can lead to slow drug development and discovery. Computational approaches, specifically machine learning (ML), to predict novel treatments and an attractive solution to this bottleneck. This dissertation seeks to address the issue of novel therapy identification using interpretable ML, specifically for the areas of drug repurposing, novel target identification and the prediction of efficacious drug combinations. To start we introduce a method, CATNIP, that identifies potential disease areas for drug classes and specific drugs. CATNIP relies solely on drug characteristics (chemical and biological) which allows it to be used for both FDA approved and investigational drugs. Additionally, CATNIP is built using easily interpreted features, which allows for potential drug-disease mechanisms to be elucidated from predictions. We demonstrate the strength of CATNIP’s predictions and present multiple drug class candidate repurposing opportunities. Secondly, we present ECLIPSE which can predict essential genes within a wide array of cancer types. This model is built using experimental loss-of-function screening results from shRNA and CRISPR. The integrated use of both platforms allows ECLIPSE to identify potential therapeutic targets better than either platform alone. Building upon our work within cancer, we present a suite of models that predict cancer type specific drug synergy. This work thoroughly compares the current top drug synergy measures (Bliss, Loewe, etc) and characterizes the different joint mechanisms of action they represent. Our models show significant predictive performance in identifying current clinically tested drug combinations, showing promise for our novel predictions. Altogether this work demonstrates vari

关键词： 近地面臭氧   机器学习算法   交叉验证   时空分布  

摘要： 近地面臭氧由于其较强的氧化性,会对人群健康、农业生产和生态系统产生损害,需要长时间尺度、全国规模、高精度、高时空分辨率的臭氧数据产品作为优化环境风险评估和管理的支撑。目前我国臭氧地面监测数据的时空覆盖度不足,化学传输模式模拟结果的不确定性较大,而统计模型方法尤其是机器学习方法不依赖排放清单数据,建模采用的监测数据、气象观测数据、土地利用数据等更易于获取,预测精度更高,更契合本研究的目的,但该方法在国内的应用目前仍然较少。近地面臭氧浓度与气象参数等因素存在较强的关联,且针对PM2.5、NO2等大气污染物浓度的统计模型取得了良好的效果,因此基于上述数据将利用统计模型开展臭氧浓度估算研究从城市区域尺度扩展到全国尺度在技术上是基本可行的。在以预测精度为首要考虑因素,兼顾模型运行速度和可解释性的条件下,机器学习方法中的XGBoost模型更加适用于本研究。因此,本研究的目的在于利用我国地面监测数据,结合气象资料、化学传输模式模拟数据和土地利用数据等资料,基于XGBoost算法开发统计模型,从而估算我国长时间尺度、高时空分辨率的O3浓度时空分布状况,并探究我国O3浓度时空变化的关联因素。本研究充分借鉴国内外研究的经验,基于“八大综合经济区”的区划方法,在各区域分别建立单年份模型和以时间外推能力为目标的模型,并对模型表现进行了评估,最终基于模型预测的结果进行了统计分析。主要的研究结论如下:(1)将化学传输模式模拟数据作为协变量纳入XGBoost模型,以十折交叉验证为模型评估方法,可以建立具备较高预测精度的单年份模型。通过将大气化学传输模式模拟结果纳入作为协变量而不是构建因变量对于模型精度和准确度的提升均有一定的作用。不同季节和区域的模型表现存在差异,秋季的效果也相对较好,北部沿海和南部沿海经济区的效果较好。(2)将化学传输模式模拟数据作为协变量纳入XGBoost模型,以年际交叉验证为模型评估方法,可以建立具备较高时间外推性能的模型。依据时间外推的结果,本研究构建的模型不会随时间间隔的增加产生预测精度上的明显下降,间接的证据也表明也适用于全国范围的预测,具有较好的平稳性和可靠性。将化学传输模式模拟结果纳入作为协变量对于提高模型时间外推的能力具有一定的帮助,一定程度上可以减少过度拟合的情况。另外,尽管采用月均值直接建模的方式可以节约时间和计算量,但在采用同样指标对比的情况下最终效果不如直接采用逐日浓度建模,而且也无法提供更为精细时间分辨率的信息。(3)在长期空间分布上,华北地区、江浙沪地区、珠三角、四川盆地、河南地区和东北地区属于污染较为严重的地区。其中,江浙沪地区、珠三角地区大部分季节污染都较为严重,而华北地区、东北地区具有明显的季节污染特征,冬季的平均浓度较低。在长期时间变化上,我国大部分地区的臭氧浓度都呈现上升的趋势,臭氧污染正在恶化和蔓延。在年内浓度变化方面,除少部分地区外,我国整体在5月到8月浓度较高,而在1月和12月浓度较低。北部沿海地区的年内幅度变化最大。在相关因素方面,日最高温度和日照时数是预测臭氧浓度最重要的两个气象参数,同时,横纵坐标和年积日在模型的重要性体现了模型中还存在着未被其他变量所解释的时空异质性。

关键词： Remote sensing   evapotranspiration   machine learning   vegetation index  

摘要： The studies in this dissertation present evaluation of and improvement to parametric and machine learning regression methods for estimating evapotranspiration from remote sensing. It includes three main parts. The first part is an assessment of parametric regression methods for obtaining evapotranspiration from vegetation index and other variables. It was found that including more variables tends to improve results, but the form of the regression formula does not make a large difference. Algorithm performance is not as good for wetland and agricultural sites as for other land cover types. Re-training of algorithms for those surface type results in some improvement. The second part consists of an evaluation of ten machine learning techniques for retrieval of evapotranspiration from surface radiation and several other variables. It is found that the best results are obtainable using all available input variables to train the bootstrap aggregation tree, random kernel, and two- and three- hidden layer neural network algorithms. Performance is again found to be weaker for wetland and agricultural surface types than for other surface types. However, separate training of the machine learning algorithms with data from those surface types does not significantly improve performance. The third part consists of further refinement to the machine learning algorithms and application of the bootstrap aggregation tree method to generate evapotranspiration maps of the continental United States for 2012. It is found that separating snow and non-snow data points improves performance. Performance for all tested algorithms was similar against the validation data set, but best for the bootstrap aggregation tree using an independent test data set. Monthly mean maps of the continental United States are generated for the drought year 2012 using the bootstrap aggregation tree. Evapotranspiration levels are lower than those shown in comparison data sets for the growing season in the eastern Un

关键词： 公路货运量   预测   机器学习   Stacking方法   降噪自编码方法  

摘要： 随着中国经济的飞速发展,城市间的联系变得愈加密切,货运行业市场变得越来越大。公路货运作为我国目前最主要的货物运输方式,运用有效的方法准确预测公路货物运量,能够帮助改善城市交通状况、提高运输效率以及促进合理调配资源。首先对国内外公路货运量预测和机器学习的研究现状进行分析和总结,结果表明在预测货运量的研究中,使用机器学习方法比使用传统方法更具有优势。然后通过定性分析方法,综合从各个方面考虑,确定了在预测建模中使用的公路货运量影响因素。同时对传统的公路货运量预测方法和机器学习方法从理论上进行了对比分析:由于城市公路货运量预测研究中相关影响因素之间存在复杂的相关性,以及统计数据中的异常值等原因,导致传统预测模型表现欠佳,将机器学习方法引入预测研究可以提高预测的精度。近年来对机器学习的深入研究表明,深度学习领域和集成学习领域的新型机器学习方法拥有更广的适用范围,且模型预测精度相比传统机器学习方法更好。本文采用了两种新型的机器学习方法进行城市公路货运量的预测,分别是集成学习领域的Stacking方法、深度学习领域的降噪自编码方法。降噪自编码降低了传统BP神经网络模型在训练中的遇到局部极值与梯度弥散的风险。此外在自编码结构训练中加入的随机噪声,可以提高公路货运预测网络的鲁棒性和泛化能力,不易出现过拟合。Stacking相比其他集成学习算法,能选择对解决具体问题已经拥有良好性能的模型作为初级学习器,从而提高了模型训练的效率;通过交叉验证的方法构造次级训练集能使得最终生成的模型具有更高的预测精度。以深圳市公路货运量预测作为实例进行了应用研究,依照上两种机器学习方法进行建模进行预测,同时以传统方法的预测结果作对比,验证新型机器学习方法的可行性和有效性,结果表明降噪自编码模型和Stacking精度良好,均优于传统预测方法。其中Stacking在预测精度和构建模型的效率方面更优秀,可运用于城市公路货运量的预测工作。本文提出两种新型机器学习方法可为其他地区的货运量预测研究提供参考与借鉴。

关键词： 一类支持向量机   孤独森林   连续约束玻尔兹曼机   局部特异因子   地球化学异常   ROC曲线   AUC值   约登指数  

摘要： 当代的矿产资源预测与定量评价是复杂高维非线性系统的建模与评价过程,通过提取和识别地质、地球物理、地球化学及遥感等多源地学观测数据中的特异信息,并以此为依据预测矿产资源的空间分布规律。单一信息来源为依据的找矿模型存在局限性,只有通过将多种来源的观测数据进行全方位的分析与综合,才能更好的完成矿产资源勘查与定量评价工作。水系沉积物测量是区域勘查地球化学研究的重要数据源,对于寻找近地表的浅隐伏的矿产资源十分有效。本文以吉林省和龙地区为实验研究区,在Sklearn的Python语言代码基础上,开发了基于一类支持向量机、孤独森林、连续约束玻尔兹曼机和局部特异因子法的多元地球化学异常识别方法程序,从1:5万水系沉积物测量数据中提取多元地球化学异常。把实验区已知矿床和矿点的空间分布位置作为“地真”数据,对多元地球化学异常识别结果进行了ROC曲线分析,绘制了四种机器学习算法多元地球化学异常识别结果的ROC曲线并计算AUC值,用来对比四种方法的多元地球化学异常识别效果。以便更好的分析基于机器学习理论的几种方法对于多元地球化学异常识别的应用存在的差异和适用范围。研究结果表明:(1)四种机器学习算法都能够有效识别多元地球化学异常,所提取的多元地球化学异常与已知矿点具有显著的空间关联性;(2)在数据处理耗时方面孤独森林算法最优,其次是一类支持向量机,耗时最长的是连续约束玻尔兹曼机;(3)四种机器学习方法中,多元地球化学异常识别效果最好的是连续约束玻尔兹曼机,其次是一类支持向量机,最差的方法是局部特异因子方法。

关键词： Logistic回归   支持向量机   随机森林   财务舞弊  

摘要： 上市公司公开发布的财务报表是使投资者了解其公司规模、经营状况、盈利潜力等综合水平并做出相关投资决策的最主要依据。过去人们认为,数据是最具有说服力的,而且如果财务报表中的数据出了差错,很容易通过较为简单的计算来发现,因而对财务报表给予了极大的信任。然而,随着市场经济的开放和现代技术手段的发展,一些上市公司为牟取私利而对财务报表进行舞弊,且舞弊手段日渐隐蔽,有时很难被及时发现,其潜在危害十分巨大。因此,亟需有效的方法来预测财务舞弊。本文旨在通过建立数学模型,针对上市公司公布的某年度财务报表数据进行分析,根据模型预测结果判断其是否存在舞弊现象。通过查阅近年来的相关文献发现,国外对财务报表的研究主要关注公司破产和公司财务危机等问题,国内研究主要关注上市公司是否会被特别处理(ST)及其是否会出现财务困境,而对财务舞弊预测的相关研究较少,因此本文对财务舞弊预测进行一定的研究和分析是很有意义的。本文将机器学习方法应用到财务舞弊的预测中。机器学习既可以自上而下的验证或反驳假设,又可以自下而上的从数据中得出无假设的结论。因此,本文采用机器学习方法分别建立了三种模型:Logistic回归模型,支持向量机(SVM)模型,以及随机森林(RF)模型。Logistic回归模型在发现隐藏的数据信息方面应用广泛,以往的研究也证明其具有良好的效果,本文以此为基础进行讨论,并对模型做出了一些改进。由于财务报表是否舞弊是一个典型的分类问题,因此采用机器学习中分类和预测更加准确的算法可能会取得更好的效果。由于获取的样本量有限,而样本维数较高,并且恰为经典的二分类问题,于是处理具有这些特性数据的支持向量机模型成为了一个很好的选择。随即,由二分类自然地联想到二叉树,从而对各个决策树分类器拟合集成效果较好的随机森林模型也成为了本文的选择。由于财务报表舞弊手段存在变化,模型若能随着时间的推移添加或删除变量以及自动选择变量,则能够更加有效地识别上市公司进行舞弊的财务报表。因此,对于每个模型,本文都利用交叉验证对参数进行选择,从而对模型进行了优化。根据2013-2018年间在证监会及其下属证监局官方网站上公布的被公开行政处罚的公司名单,本文收集了舞弊公司在不同舞弊年度的财务报表数据及相应年度非舞弊上市公司的财务报表数据。所获取数据中的一部分用来建立模型,另一部分用来检测模型。由于舞弊公司在上市公司总体中的数量相对较少,本文采用不同的数据处理方法,分别建立了非平衡数据加权模型、过采样模型和欠采样模型。对于模型效果的判断,本文选取了五个指标来进行分析。结果表明,欠采样方法下的支持向量机模型的查全率最高,而欠采样方法下的随机森林模型在其他指标上表现更好,因此本文建议在基于支持向量机模型对公司是否舞弊进行预测时,结合随机森林模型进行综合考虑。最后,对模型的应用进行拓展,用模型选择出舞弊概率较大的公司,并将其从股票池中剔除,回测结果表明投资组合的收益将得到提高,表明本文的研究具有很好的应用价值。