蛋白质亚细胞定位预测研究综述

蛋白质亚细胞定位预测对于确定蛋白质功能、揭示分子交互机理、理解复杂生理过程和设计药物靶标等方面都有很大的促进作用。随着后基因组时代中蛋白质序列数据的指数增长, 研究基于机器学习的计算性蛋白质亚细胞定位预测方法变得越来越重要。为了能够把握该问题的研究状况, 从数据集构建、蛋白质特征提取与表示、预测算法设计、算法测试和Web服务的建立等五个方面对蛋白质亚细胞定位预测的研究进行了综述。指出了目前该研究领域需要解决的核心问题及难点问题, 分析了当前研究中出现的一些新情况, 并对将来的研究方向和研究重点进行了展望。     

一种新的蛋白质亚细胞定位预测方法

蛋白质亚细胞定位是蛋白质组学基本问题之一。某些类型蛋白质可能存在于两个或两个以上的亚细胞位置,这类蛋白质的亚细胞定位问题更为复杂。分别利用Gene Ontology和伪氨基酸成分法,将一条蛋白质表示为一实值向量;采纳多标记学习中的Ranking思想,计算出一得分向量V,该向量的每一分量的值表示被预测蛋白质属于某个亚细胞位置的概率;利用最近邻算法预测蛋白质所属亚细胞位置的个数n,得分向量V中得分最高的n个分量对应的亚细胞位置即为预测的位置。     

Towards a better prediction of subcellular location of long non-coding RNA

The spatial distribution pattern of long non-coding RNA (lncRNA) in cell is tightly related to their function. With the increment of publicly available subcellular location data, a number of computational methods have been developed for the recognition of the subcellular localization of lncRNA. Unfortunately, these computational methods suffer from the low discriminative power of redundant features or overfitting of oversampling. To address those issues and enhance the prediction performance, we present a support vector machine-based approach by incorporating mutual information algorithm and incremental feature selection strategy. As a result, the new predictor could achieve the overall accuracy of 91.60%. The highly automated web-tool is available at lin-group.cn/server/iLoc-LncRNA(2.0)/website. It will help to get the knowledge of lncRNA subcellular localization.     

基于三层集成多标记学习的蛋白质多亚细胞定位预测

针对多标记学习和集成学习在解决蛋白质多亚细胞定位预测问题上应用还不成熟的状况,研究基于集成多标记学习的蛋白质多亚细胞定位预测方法。首先,从多标记学习和集成学习相结合的角度提出了一种三层的集成多标记学习系统框架结构,该框架将学习算法和分类器进行了层次性分类,并把二分类学习、多分类学习、多标记学习和集成学习进行有效整合,形成一个通用型的三层集成多标记学习模型;其次,基于面向对象技术和统一建模语言（UML）对系统模型进行了设计,使系统具备良好的可扩展性,通过扩展手段增强系统的功能和提高系统的性能;最后,使用Java编程技术对模型进行扩展,实现了一个学习系统软件,并成功应用于蛋白质多亚细胞定位预测问题上。通过在革兰氏阳性细菌数据集上进行测试,验证了系统功能的可操作性和较好的预测性能,该系统可以作为解决蛋白质多亚细胞定位预测问题的一个有效工具。     

使用伪氨基酸和集成分类器预测凋谢蛋白亚细胞定位

预测凋谢蛋白质亚细胞定位是生物信息学和蛋白质科学中重要的研究内容.基于Chou的伪氨基酸组成概念,用近似熵表示蛋白质序列的附加特征,组成新的伪氨基酸组成表示序列特征.将蛋白质序列看作短时间序列,近似熵能够区分不同亚细胞定位中序列的复杂度.结合多个模糊K近邻分类器(基本分类器)的集成分类器作为预测工具.以不同维数的伪氨基酸组成向量,作为每个基本分类器的输入数据.3个常用的数据集用来测试算法的性能,Jackknife测试结果表明新算法有效和实用.有望发展成为亚细胞定位研究的有用工具.     

Evolutions in machine learning technology for financial distress prediction: A comprehensive review and comparative analysis

In recent years, financial distress prediction (FDP), also known as corporate failure prediction or bankruptcy prediction, has gained significant importance due to its impact on organizations, especially during unexpected events like pandemics and wars. Machine learning (ML) models have emerged as innovative and essential tools in predicting financial distress, leveraging the ever-increasing volume of databases and computing power. This study utilizes bibliographic techniques to contribute to the field's literature review to address the disorganized nature of the existing literature on FDP, reduce confusion, and provide clarity to domain researchers. These techniques enable identifying the progress of articles published over the years, influential authors, and highly cited articles. Additionally, the study examines crucial aspects of data preprocessing, such as missing data, imbalanced data, feature selection, and outliers, as they significantly impact the robustness and performance of ML models. Furthermore, it discusses essential models employed in FDP, focusing on recent advancements that represent promising trends. In conclusion, this study contributes to the field by uncovering novel trends and proposing possible directions for advancing FDP research. These findings will guide researchers, practitioners, and stakeholders in their quest for improved prediction and decision-making in financial distress.     

深度学习人脸特征点自动定位综述

人脸特征点定位是根据输入的人脸数据自动定位出预先按人脸生理特征定义的眼角、鼻尖、嘴角和脸部轮廓等面部关键特征点,在人脸识别和分析等系统中起着至关重要的作用。本文对基于深度学习的人脸特征点自动定位进行综述,阐释了人脸特征点自动定位的含义,归纳了目前常用的人脸公开数据集,系统阐述了针对2维和3维数据特征点的自动定位方法,总结了各方法的研究现状及其应用,分析了当前人脸特征点自动定位技术在深度学习应用中的现状、存在问题及发展趋势。在公开的2维和3维人脸数据集上对不同方法进行了比较。通过研究可以看出,基于深度学习的2维人脸特征点的自动定位方法研究相对比较深入,而3维人脸特征点定位方法的研究在模型表示、处理方法和样本数量上都存在挑战。未来基于深度学习的3维人脸特征点定位方法将成为研究趋势。     

蛋白质结构预测的拟物拟人算法研究*

根据国际上最新提出的蛋白质结构预测问题的三维欧氏空间连续模型,找到了相应的物理模型,并且设计了适当的拟人策略,形成了相应的拟物拟人算法.     

Text prediction systems: a survey

Text prediction is one of the most widely used techniques to enhance the communication rate in augmentative and alternative communication. Prediction systems are traditionally used by people with disabilities (e.g. people with motor and speech impairments). However, new applications, such as writing short text messages via mobile phones, have recently appeared. A vast amount of heterogeneous text prediction methods and techniques can be found in literature. Their heterogeneity makes it difficult to understand and compare them, in order to select the most convenient technique for a specific design. This paper presents a survey on text prediction techniques with the intention to provide a systematic view of this field. Prediction applications and related features, such as block size, dictionary structure, prediction method, user interface, etc., are examined. In addition, prediction measurement parameters and published results are compared. A large number of factors that may influence prediction results, including the acceptance of the system by the users, are reviewed, and their influence on the performance and usability of the system is discussed.

Spray and forward:Efficient routing based on the Markov location prediction model for DTNs

Typical delay tolerant networks(DTNs)often suffer from long and variable delays,frequent connectivity disruptions,and high bit error rates.In DTNs,the design of an efficient routing algorithm is one of the key issues.The existing methods improve the accessibility probability of the data transmission by transmitting many copies of the packet to the network,but they may cause a high network overhead.To address the tradeoff between a successful delivery ratio and the network overhead,we propose a DTN routing algorithm based on the Markov location prediction model,called the spray and forward routing algorithm(SFR).Based on historical information of the nodes,the algorithm uses the second-order Markov forecasting mechanism to predict the location of the destination node,and then forwards the data by greedy routing,which reduces the copies of packets by spraying the packets in a particular direction.In contrast to a fixed mode where a successful-delivery ratio and routing overhead are contradictory,a hybrid strategy with multi-copy forwarding is able to reduce the copies of the packets efficiently and at the same time maintain an acceptable successful-delivery ratio.The simulation results show that the proposed SFR is efficient enough to provide better network performance than the spray and wait routing algorithm,in scenarios with sparse node density and fast mobility of the nodes.     

The prediction of protein secondary structure with a Cascade Correlation Learning Architecture of neural networks

A Cascade Correlation Learning Architecture (CCLA) of neural networks is tested on the task of predicting the secondary structure of proteins. The results are compared with those obtained with Neural Networks (NN) trained with the back-propagation algorithm (BPNN) and generated with genetic algorithms. CCLA proceeds towards the global minimum of the error function more efficiently than BPNN. However, only a slight improvement in the average efficiency value is noticeable (61.82% as compared with 61.61% obtained with BPNN). The values of the three correlation coefficients for the discriminated secondary structures are also rather similar (C_t8,C ,C and C_coil are 0.36, 0.29 and 0.36 with CCLA, and 0.36, 0.31 and 0.35 with BPNN). This indicates that the efficiency of the prediction does not depend upon the training algorithm, and confirms our previous observation that when single sequences are used as input code to the network system, different NN architectures can perform similarly.     

在线社交网络中用户地理位置预测综述

随着智能移动终端的日益普及,人们越来越多地利用社交网络平台（如Twitter、新浪微博等）获取信息、评论和交流. 虽然全球卫星定位系统（GPS）设备能够精确获取位置信息,但是大量用户出于隐私和安全的考虑不会直接共享自己的位置信息. 因此,如何获取在线用户的地理位置成为了一个前沿的研究领域以及学术界和工业界共同关注的重要课题,并且成为众多下游应用的基础,例如基于位置的定向广告投放、事件/地点的推荐、自然灾害或疾病预警和网络犯罪的追踪等. 详细总结了预测社交网络用户地理位置的方法、数据、评价体系和基础算法. 首先,归纳了不同的定位任务以及相应的评价指标;其次,针对不同的任务梳理所用的数据类型和数据融合方式,并且,详尽分析了已有的信息抽取和特征选择方式及其优缺点;再次,对现有定位模型和算法进行分类,从地名词典、传统机器学习和深度学习3个方面对用户定位方法进行阐述和分析;最后,总结了社交网络用户地理位置预测的难点和面临的挑战,并展望该领域的发展趋势和未来研究所需要关注的方向.

     

PITAGORAS-PSP: Including domain knowledge in a multi-objective approach for protein structure prediction

Whenever evolutionary algorithms are used to solve certain classes of problems such as those that present a huge search space, the incorporation of problem-specific knowledge is required to achieve adequate levels of performance. In this paper, we propose a multi-objective optimization-based procedure that includes such a domain-specific knowledge to cope with a difficult problem, the protein structure prediction (PSP). This problem is considered to be an open problem as there is no recognized “best” procedure to find solutions. It presents a vast search space and the analysis of each protein conformation requires significant amount of computing time. In our procedure, we provide a reduction of the search space by using the dependent rotamer library and include new heuristics to improve a multi-objective approach to PSP based on the PAES algorithm. As it is shown in the paper, by using benchmark proteins from the CASP8 set, this hybrid PSP procedure provides competitive results when it is compared with some of the better proposals appeared up to now.     

静态软件缺陷预测方法研究

开源软件缺陷预测通过挖掘软件历史仓库的数据,利用与软件缺陷相关的度量元或源代码本身的语法语义特征,借助机器学习或深度学习方法提前发现软件缺陷,从而减少软件修复成本并提高产品质量. 漏洞预测则通过挖掘软件实例存储库来提取和标记代码模块,预测新的代码实例是否含有漏洞,减少漏洞发现和修复的成本. 通过对2000年至2022年12月软件缺陷预测研究领域的相关文献调研,以机器学习和深度学习为切入点,梳理了基于软件度量和基于语法语义的预测模型. 基于这2类模型,分析了软件缺陷预测和漏洞预测之间的区别和联系,并针对数据集来源与处理、代码向量的表征方法、预训练模型的提高、深度学习模型的探索、细粒度预测技术、软件缺陷预测和漏洞预测模型迁移六大前沿热点问题进行了详尽分析,最后指出了软件缺陷预测未来的发展方向.

     

Context-prediction performance by a dynamic Bayesian network: Emphasis on location prediction in ubiquitous decision support environment

Ubiquitous decision support systems require more intelligent mechanism in which more timely and accurate decision support is available. However, conventional context-aware systems, which have been popular in the ubiquitous decision support systems field, cannot provide such agile and proactive decision support. To fill this research void, this paper proposes a new concept of context prediction mechanism by which the ubiquitous decision support devices are able to predict users’ future contexts in advance, and provide more timely and proactive decision support that users would be satisfied much more. Especially, location prediction is useful because ubiquitous decision support systems could dynamically adapt their decision support contents for a user based on a user’s future location. In this sense, as an alternative for the inference engine mechanism to be used in the ubiquitous decision support systems capable of context-prediction, we propose an inductive approach to recognizing a user’s location by learning a dynamic Bayesian network model. The dynamic Bayesian network model has been evaluated with a set of contextual data from undergraduate students. The evaluation result suggests that a dynamic Bayesian network model offers significant predictive power in the location prediction. Besides, we found that the dynamic Bayesian network model has a great potential for the future types of ubiquitous decision support systems.     

Interest rate prediction: a neuro-hybrid approach with data preprocessing

The following research implements a differential evolution-based fuzzy-type clustering method with a fuzzy inference neural network after input preprocessing with regression analysis in order to predict future interest rates, particularly 3-month T-bill rates. The empirical results of the proposed model is compared against nonparametric models, such as locally weighted regression and least squares support vector machines, along with two linear benchmark models, the autoregressive model and the random walk model. The root mean square error is reported for comparison.     

Optimal pressure sensor placement and assessment for leak location using a relaxed isolation index: Application to the Barcelona water network

Water distribution networks are large complex systems affected by leaks, which often entail high costs and may severely jeopardise the overall water distribution performance. Successful leak location is paramount in order to minimize the impact of these leaks when occurring. Sensor placement is a key issue in the leak location process, since the overall performance and success of this process highly depends on the choice of the sensors gathering data from the network. Common problems when isolating leaks in large scale highly gridded real water distribution networks include leak mislabelling and the obtention of large number of possible leak locations. This is due to similarity of leak effect in the measurements, which may be caused by topological issues and led to incomplete coverage of the whole network. The sensor placement strategy may minimize these undesired effects by setting the sensor placement optimisation problem with the appropriate assumptions (e.g. geographically cluster alike leak behaviors) and by taking into account real aspects of the practical application, such as the acceptable leak location distance. In this paper, a sensor placement methodology considering these aspects and a general sensor distribution assessment method for leak diagnosis in water distribution systems is presented and exemplified with a small illustrative case study. Finally, the proposed method is applied to two real District Metered Areas (DMAs) located within the Barcelona water distribution network.     

Online set-point optimisation cooperating with predictive control of a yeast fermentation process: A neural network approach

Online set-point optimisation which cooperates with model predictive control (MPC) and its application to a yeast fermentation process are described. A computationally efficient multilayer control system structure with adaptive steady-state target optimisation (ASSTO) and a suboptimal MPC algorithm are presented in which two neural models of the process are used. For set-point optimisation, a steady-state neural model is linearised online and the set-point is calculated from a linear programming problem. For MPC, a dynamic neural model is linearised online and the control policy is calculated from a quadratic programming problem. In consequence of linearisation of neural models, the necessity of online nonlinear optimisation is eliminated. Results obtained in the proposed structure are comparable with those achieved in a computationally demanding structure with nonlinear optimisation used for set-point optimisation and MPC.