Instance-based Machine Learning Research Articles

Predicting micro/nanoscale colloidal interactions through local neighborhood graph neural networks

Understanding interparticle interactions has been one of the most important topics of research in the field of micro/nanoscale materials. Many significant characteristics of such materials directly stem from the way their building blocks interact with each other. In this work, we investigate the efficacy of a specific category of Machine Learning (ML) methods known as interaction networks in predicting interparticle interactions within colloidal systems. We introduce and study Local Neighborhood Graph Neural Networks (LN-GNNs), defined according to the local environment of colloidal particles derived from particle trajectory data. The LN-GNN framework is trained for unique categories of particle neighborhood environments in order to predict interparticle interactions. We compare the performance of the LN-GNN to a baseline interaction network with a simpler architecture and to an Instance-Based ML algorithm, which is computationally more expensive. We find that the prediction performance of LN-GNN measured as an average normalized mean absolute error outperforms the baseline interaction network by a factor of 2–10 for different local neighborhood configurations. Furthermore, LN-GNN’s performance turns out to be very comparable to the instance-based ML framework while being an order of magnitude less expensive in terms of the required computation time. The results of this work can provide the foundations for establishing accurate models of colloidal particle interactions that are derived from real particle trajectory data.

Detecting False Data Injection Attacks in Peer to Peer Energy Trading Using Machine Learning

In peer-to-peer (P2P) energy trading, the incorporation of distributed energy resources with unprotected data, originating from sources such as home energy management systems that are connected through the Internet, provokes vulnerabilities that can manifest security breaches. In this article, two threat scenarios based on a novel false data injection attack (FDIA) model in a local P2P energy trading system are explored. In these scenarios, an attacker gains free energy by manipulating prosumers’ consumption and demand. Precise and fast attack detection is needed to guarantee suitable countermeasures to prevent potential risks. We propose a novel instance-based machine learning (ML) classifier for detecting FDIAs. In contrast to black-box ML models, our algorithm provides a transparent decision-making procedure with significant predictive performance. We apply our detection model to a real-world dataset from Austin, Texas. Our experimental results show superior performance as compared to several popular interpretable and non-interpretable ML methods. On average, we achieve a 96.10 percent detection rate, a 96.18 percent accuracy rate, and a false negative rate of 1.97 percent with our approach.

Detecting block ciphers generic attacks: An instance-based machine learning method

Cryptography facilitates selective communication through encryption of messages and or data. Block-cipher processing is one of the prominent methods for modern cryptographic symmetric encryption schemes. The rise in attacks on block-ciphers led to the development of more difficult encryption schemes. However, attackers decrypt block-ciphers through generic attacks given sufficient time and computing. Recent research had applied machine learning classification algorithms to develop intrusion detection systems to detect multiple types of attacks. These intrusion detection systems are limited by misclassifying generic attacks and suffer reduced effectiveness when evaluated for detecting generic attacks only. Hence, this study introduced and proposed k-nearest neighbors, an instance-based machine learning classification algorithm, for the detection of generic attacks on block-ciphers. The value of k was varied (i.e., 1, 3, 5, 7, and 9) and multiple nearest neighbors classification models were developed and evaluated using two distance functions (i.e., Manhattan and Euclidean) for classifying between generic attacks and normal network packets. All nearest neighbors models using the Manhattan distance function performed better than their Euclidean counterparts. The 1-nearest neighbor (Manhattan distance function) model had the highest overall accuracy of 99.6%, a generic attack detection rate of 99.5% which tallies with the 5, 7, and 9 nearest neighbors models, and a false alarm rate of 0.0003 which is the same for all Manhattan nearest neighbors classification models. These instance-based methods performed better than some existing methods that even implemented an ensemble of deep-learning algorithms. Therefore, an instance-based method is recommended for detecting block-ciphers generic attacks.

Model-free data-driven computational mechanics enhanced by tensor voting

The data-driven computing paradigm initially introduced by Kirchdoerfer and Ortiz (2016) is extended by incorporating locally linear tangent spaces into the data set. These tangent spaces are constructed by means of the tensor voting method introduced by Mordohai and Medioni (2010) which improves the learning of the underlying structure of a data set. Tensor voting is an instance-based machine learning technique which accumulates votes from the nearest neighbors to build up second-order tensors encoding tangents and normals to the underlying data structure. The here proposed second-order data-driven paradigm is a plug-in method for distance-minimizing as well as entropy-maximizing data-driven schemes. Like its predecessor (Kirchdoerfer and Ortiz, 2016), the resulting method aims to minimize a suitably defined free energy over phase space subject to compatibility and equilibrium constraints. The method’s implementation is straightforward and numerically efficient since the data structure analysis is performed in an offline step. Selected numerical examples are presented that establish the higher-order convergence properties of the data-driven solvers enhanced by tensor voting for ideal and noisy data sets.

Integrating Machine Learning With Microsimulation to Classify Hypothetical, Novel Patients for Predicting Pregabalin Treatment Response Based on Observational and Randomized Data in Patients With Painful Diabetic Peripheral Neuropathy.

PurposeVariability in patient treatment responses can be a barrier to effective care. Utilization of available patient databases may improve the prediction of treatment responses. We evaluated machine learning methods to predict novel, individual patient responses to pregabalin for painful diabetic peripheral neuropathy, utilizing an agent-based modeling and simulation platform that integrates real-world observational study (OS) data and randomized clinical trial (RCT) data.Patients and methodsThe best supervised machine learning methods were selected (through literature review) and combined in a novel way for aligning patients with relevant subgroups that best enable prediction of pregabalin responses. Data were derived from a German OS of pregabalin (N=2642) and nine international RCTs (N=1320). Coarsened exact matching of OS and RCT patients was used and a hierarchical cluster analysis was implemented. We tested which machine learning methods would best align candidate patients with specific clusters that predict their pain scores over time. Cluster alignments would trigger assignments of cluster-specific time-series regressions with lagged variables as inputs in order to simulate “virtual” patients and generate 1000 trajectory variations for given novel patients.ResultsInstance-based machine learning methods (k-nearest neighbor, supervised fuzzy c-means) were selected for quantitative analyses. Each method alone correctly classified 56.7% and 39.1% of patients, respectively. An “ensemble method” (combining both methods) correctly classified 98.4% and 95.9% of patients in the training and testing datasets, respectively.ConclusionAn ensemble combination of two instance-based machine learning techniques best accommodated different data types (dichotomous, categorical, continuous) and performed better than either technique alone in assigning novel patients to subgroups for predicting treatment outcomes using microsimulation. Assignment of novel patients to a cluster of similar patients has the potential to improve prediction of patient outcomes for chronic conditions in which initial treatment response can be incorporated using microsimulation.Clinical trial registrieswww.clinicaltrials.gov: NCT00156078, NCT00159679, NCT00143156, NCT00553475.

Significance of Artificial Intelligence and Machine Learning Techniques in Smart Cloud Computing: A Review

Realization of the tremendous features and facilities provided by Cloud Computing by the geniuses in the world of digital marketing increases its demand. As customer satisfaction is the manifest of this ever shining field, balancing its load becomes a major issue. Various heuristic and meta-heuristic algorithms were applied to get optimum solutions. The current era is much attracted with the provisioning of self-manageable, self-learnable, self-healable, and self-configurable smart systems. To get self-manageable Smart Cloud, various Artificial Intelligence and Machine Learning (AI-ML) techniques and algorithms are revived. In this review, recent trend in the utilization of AI-ML techniques, their applied areas, purpose, their merits and demerits are highlighted. These techniques are further categorized as instance-based machine learning algorithms and reinforcement learning techniques based on their ability of learning. Reinforcement learning is preferred when there is no training data set. It leads the system to learn by its own experience itself even in dynamic environment.

Constraint nearest neighbor for instance reduction

In instance-based machine learning, algorithms often suffer from prohibitive computational costs and storage space. To overcome such problems, various instance reduction techniques have been developed to remove noises and/or redundant instances. Condensation approach is the most frequently used method, and it aims to remove the instances far away from the decision surface. Edition method is another popular one, and it removes noises to improve the classification accuracy. Drawbacks of these existing techniques include parameter dependency and relatively low accuracy and reduction rate. To solve these drawbacks, the constraint nearest neighbor-based instance reduction (CNNIR) algorithm is proposed in this paper. We firstly introduce the concept of natural neighbor and apply it into instance reduction to eliminate noises and search core instances. Then, we define a constraint nearest neighbor chain which only consists of three instances. It is used to select border instances which can construct a rough decision boundary. After that, a specific strategy is given to reduce the border set. Finally, reduced set is obtained by merging border and core instances. Experimental results show that compared with existing algorithms, the proposed algorithm effectively reduces the number of instances and achieves higher classification accuracy. Moreover, it does not require any user-defined parameters.

Fast Data Reduction With Granulation-Based Instances Importance Labeling

Data reduction has become greatly significant prior to applying instance-based machine learning algorithms in the Big Data era. Data reduction is used to reduce the size of data sets while retaining representative data. Existing algorithms, however, suffer from heavy computational cost and in having tradeoff in size reduction rate and learning accuracy. In this paper, we propose a fast data reduction approach by using granular computing to label important instances, i.e., instances with higher contributions to the learning task. The original data set is first granulated into $K$ granules by applying $K$ -means to a mapped lower-dimension space. Then, the importance of each instance in every granule is labeled based on its Hausdorff distance. Those instances whose importance values are lower than an experimentally tuned threshold are eliminated. The presented algorithm is applied to $k$ NN classification tasks with eighteen different sizes of data sets from the UCI repository, and its outstanding performance in classification accuracy, size reduction rate, and runtime is illustrated by comparing with seven data reduction methods. The experimental results demonstrate that the proposed algorithm can greatly reduce the computational cost and achieve a higher classification accuracy when the reduction size is the same for all the compared algorithms.

Using time series analysis approaches for improved prediction of pain outcomes in subgroups of patients with painful diabetic peripheral neuropathy.

Prior work applied hierarchical clustering, coarsened exact matching (CEM), time series regressions with lagged variables as inputs, and microsimulation to data from three randomized clinical trials (RCTs) and a large German observational study (OS) to predict pregabalin pain reduction outcomes for patients with painful diabetic peripheral neuropathy. Here, data were added from six RCTs to reduce covariate bias of the same OS and improve accuracy and/or increase the variety of patients for pain response prediction. Using hierarchical cluster analysis and CEM, a matched dataset was created from the OS (N = 2642) and nine total RCTs (N = 1320). Using a maximum likelihood method, we estimated weekly pain scores for pregabalin-treated patients for each cluster (matched dataset); the models were validated with RCT data that did not match with OS data. We predicted novel ‘virtual’ patient pain scores over time using simulations including instance-based machine learning techniques to assign novel patients to a cluster, then applying cluster-specific regressions to predict pain response trajectories. Six clusters were identified according to baseline variables (gender, age, insulin use, body mass index, depression history, pregabalin monotherapy, prior gabapentin, pain score, and pain-related sleep interference score). CEM yielded 1766 patients (matched dataset) having lower covariate imbalances. Regression models for pain performed well (adjusted R-squared 0.90–0.93; root mean square errors 0.41–0.48). Simulations showed positive predictive values for achieving >50% and >30% change-from-baseline pain score improvements (range 68.6–83.8% and 86.5–93.9%, respectively). Using more RCTs (nine vs. the earlier three) enabled matching of 46.7% more patients in the OS dataset, with substantially reduced global imbalance vs. not matching. This larger RCT pool covered 66.8% of possible patient characteristic combinations (vs. 25.0% with three original RCTs) and made prediction possible for a broader spectrum of patients.Trial Registration: www.clinicaltrials.gov (as applicable): NCT00156078, NCT00159679, NCT00143156, NCT00553475.

Refinement operators for directed labeled graphs with applications to instance-based learning

This paper presents a collection of refinement operators for directed labeled graphs (DLGs), and a family of distance and similarity measures based on them. We build upon previous work on refinement operators for other representations such as feature terms and description logic models. Specifically, we present eight refinement operators for DLGs, which will allow for the adaptation of three similarity measures to DLGs: the anti-unification-based, Sλ, the property-based, Sπ, and the weighted property-based, Swπ, similarities. We evaluate the resulting measures empirically, comparing them to existing similarity measures for structured data in the context of instance-based machine learning.

Modeling the Impact of Urban Landscape Change on Urban Wetlands Using Similarity Weighted Instance-Based Machine Learning and Markov Model

Urban wetlands play important roles in providing several ecosystem services that support the urban environment. As such, scientists have studied them to understand the urban processes that lead to their continued decline. However, little attention has been given to the drivers of land-use change that may affect this fragile ecosystem in the future. Understanding this could serve as a critical step towards urban wetland management and sustainability. In this study, we utilized an integrated approach that combined Similarity Weighted Instance-based Machine Learning and Markov chain, both embedded in the IDRISI Land Change Modeler to simulate change in the landscape of three watersheds in the Kansas City Metropolitan area. The purpose was to assess the possible future impacts of urban expansion-induced landscape change on wetlands within the study area, using a retrospective approach. To achieve this, classified SPOT satellite data covering the three watersheds were used to generate historical land cover maps of the study area between 1992 and 2010 to analyze changes to the landscape. In addition, the study identified several drivers of land change associated with the historical change process in the study area, and accounted for their role in the modeling process. On this basis, the study made the prediction of urban landscape transformation to the end date of 2014. The prediction result was verified with a more accurate map that was derived from independently classifying a 2014 SPOT image of the study area. Results from this study show that impervious surfaces, which were used as an index of urban expansion, may increase by approximately the same magnitude experienced historically, which may result in a small but significant loss of wetlands and other land cover classes within the study area.

Integration of Markov chain analysis and similarity-weighted instance-based machine learning algorithm (SimWeight) to simulate urban expansion

ABSTRACTThis study simulates urban expansion using Kaduna in North-West Nigeria as a case study. A hybrid model that integrates the similarity-weighted instance-based machine learning algorithm for transition potential modelling and the Markov chain model to quantify and allocate land-use change was used to overcome the identified weaknesses of known modelling techniques such as the cellular automata, Markov chain and standard logistic regression models. Environmental and urban physical variables that act as constraints and/or incentives to urban expansion were operationalized to create transition potentials for spatiotemporal states of built-up land use for the year 1990 and 2001. Model evaluation and validation was carried out using the relative operating characteristic and kappa index of agreement statistics. Having obtained satisfactory outcomes from the validation process, the modelled transition potentials were used to predict future urban expansion for forthcoming years. The simulated land-use maps provide valuable insights into the location and type of urban expansion that is likely to occur in Kaduna in the foreseeable future. This provides city managers and planners much needed information that could inform urban policy aimed at better planning and management of urban development.

An Online Rolling Optimal Control Strategy for Commuter Hybrid Electric Vehicles Based on Driving Condition Learning and Prediction

This paper presents an online supervisory control strategy for commuter hybrid electric vehicles (HEVs) based on driving condition learning and prediction. The aim is to provide an online self-learning-based framework to keep Pontryagin's minimum principle (PMP)-based control adapting to the time-varying driving condition on commuting routes and minimize fuel consumption. There are three steps to realizing this strategy. First, two novel statistical features are proposed to describe the frequency distribution of achievable working points of the hybrid powertrain under a driving condition. Second, based on the characteristic that commuting trips with similar trip start time, direction, and weather condition have similar driving conditions, we develop an instance-based machine learning algorithm to learn the driving condition. A k -nearest neighbor ( k -NN) prediction algorithm is used to predict future driving conditions. Third, we establish an online supervisory control strategy, together with rolling driving condition prediction and optimal costate approximation. The approximation algorithm can approximate the optimal constant costate for the entire prediction horizon just based on the proposed two features. Simulation study and bench tests are conducted on a parallel hybrid powertrain of a city bus using standard driving cycles and real-world sampled commuting trips with different trip start times and weather conditions. The results show that the proposed driving condition learning and prediction method is effective for predicting the upcoming driving conditions. Meanwhile, the proposed strategy shows substantial improvement of fuel economy compared with the rule-based (RB) strategy and the adaptive equivalent consumption minimization strategy (A-ECMS).

Inducing models of behavior from expert task performance in virtual environments

We developed an end-to-end process for inducing models of behavior from expert task performance through in-depth case study. A subject matter expert (SME) performed navigational and adversarial tasks in a virtual tank combat simulation, using the dTank and Unreal platforms. Using eye tracking and Cognitive Task Analysis, we identified the key goals pursued by and attributes used by the SME, including reliance on an egocentric spatial representation, and on the fly re-representation of terrain in qualitative terms such as “safe” and “risky”. We demonstrated methods for automatic extraction of these qualitative higher-order features from combinations of surface features present in the simulation, producing a terrain map that was visually similar to the SME annotated map. The application of decision-tree and instance-based machine learning methods to the transformed task data supported prediction of SME task selection with greater than 95 % accuracy, and SME action selection at a frequency of 10 Hz with greater than 63 % accuracy, with real time constraints placing limits on algorithm selection. A complete processing model is presented for a path driving task, with the induced generative model deviating from the SME chosen path by less than 2 meters on average. The derived attributes also enabled environment portability, with path driving models induced from dTank performance and deployed in Unreal demonstrating equivalent accuracy to those induced and deployed completely within Unreal.

Unsupervised Conversation Extraction in Short Text Message Streams

PDF HTML阅读 XML下载 导出引用 引用提醒 短文本信息流的无监督会话抽取技术 DOI: 10.3724/SP.J.1001.2012.04031 作者: 作者单位: 作者简介: 通讯作者: 中图分类号: 基金项目: 国家自然科学基金(60933005, 60873204); 国家高技术研究发展计划(863)(2001AA012505); 国家242 信息安全计划课题(2009A90) Unsupervised Conversation Extraction in Short Text Message Streams Author: Affiliation: Fund Project: 摘要 | 图/表 | 访问统计 | 参考文献 | 相似文献 | 引证文献 | 资源附件 | 文章评论 摘要:文本会话抽取将网络聊天记录等短文本信息流中的信息根据其所属的会话分检到多个会话队列,有利于短文本信息的管理及进一步的挖掘.现有的会话抽取技术主要对基于文本相似度的聚类方法进行改进,面临着短文本信息流的特征稀疏性、奇异性和动态性等挑战.针对这些挑战,研究无监督的会话抽取技术,提出了一种基于信息流时序特征和上下文相关度的抽取方法.首先研究了信息流的会话生命周期规律,提出基于信息产生频率的会话边界检测方法;其次提出信息间的上下文相关度概念,采用基于实例的机器学习方法计算该相关度;最后综合信息产生频率和上下文相关度,设计了基于Single-Pass 聚类模型的会话在线抽取算法SPFC(single-pass based on frequency and correlation).真实数据集上的实验结果表明,SPFC算法与已有的基于文本相似度的会话抽取算法相比,F1 评测指标提高了30%. Abstract:Short text message streams are produced by Short Message Service, Instant Messager and BBS, which are widely used. Each stream usually contains. Extracting the conversations in the streams is helpful to various applications including business intelligence, investigation of crime and public opinion analysis. Existing research mainly based on text similarity encounter challenges such as the anomaly, dynamics, and the sparse eigenvector of short text message. This paper proposes an innovative conversation extraction method to cover the challenges. Firstly, the study detects the conversation boundary of short text message streams using temporal feature; secondly, contextually correlative degree is introduced to replace similar degree, and an instance-based machine learning method is proposed to compute the correlative degree. Finally, the study designs Single-Pass based conversation extraction algorithm SPFC (single-pass based on frequency and correlation), which combines the temporal and contextually correlative characteristics. Experimental results on a large real Chinese dataset show that this method SPFC improves the performance by 30% when compared with the best existing variation algorithm in terms of F1 measure. 参考文献 相似文献 引证文献

A class boundary preserving algorithm for data condensation

In instance-based machine learning, algorithms often suffer from storing large numbers of training instances. This results in large computer memory usage, long response time, and often oversensitivity to noise. In order to overcome such problems, various instance reduction algorithms have been developed to remove noisy and surplus instances. This paper discusses existing algorithms in the field of instance selection and abstraction, and introduces a new approach, the Class Boundary Preserving Algorithm (CBP), which is a multi-stage method for pruning the training set, based on a simple but very effective heuristic for instance removal. CBP is tested with a large number of datasets and comparatively evaluated against eight of the most successful instance-based condensation algorithms. Experiments showed that our algorithm achieved similar classification accuracies, with much improved storage reduction and competitive execution speeds.

Log integration on large scale for global networking monitoring

Supposing that the overall situation is dug out from the distributed monitoring nodes, there should be two critical obstacles, heterogenous schema and instance, to integrating heterogeneous data from different monitoring sensors. To tackle the challenge of heterogenous schema, an instance-based approach for schema mapping, named instance-based machine-learning (IML) approach was described. And to solve the problem of heterogenous instance, a novel approach, called statistic-based clustering (SBC) approach, which utilized clustering and statistics technologies to match large scale sources holistically, was also proposed. These two algorithms utilized the machine-leaning and clustering technology to improve the accuracy. Experimental analysis shows that the IML approach is more precise than SBC approach, reaching at least precision of 81% and recall rate of 82%. Simulation studies further show that SBC can tackle large scale sources holistically with 85% recall rate when there are 38 data sources.

Horse breed discrimination using machine learning methods

Genetic relationships and population structure of 8 horse breeds in the Czech and Slovak Republics were investigated using classification methods for breed discrimination. To demonstrate genetic differences among these breeds, we used genetic information - genotype data of microsatellite markers and classification algorithms - to perform a probabilistic prediction of an individual's breed. In total, 932 unrelated animals were genotyped for 17 microsatellite markers recommended by the ISAG for parentage testing (AHT4, AHT5, ASB2, HMS3, HMS6, HMS7, HTG4, HTG10, VHL20, HTG6, HMS2, HTG7, ASB17, ASB23, CA425, HMS1, LEX3). Algorithms of classification methods - J48 (decision trees); Naive Bayes, Bayes Net (probability predictors); IB1, IB5 (instance-based machine learning methods); and JRip (decision rules) - were used for analysis of their classification performance and of results of classification on this genotype dataset. Selected classification methods (Naive Bayes, Bayes Net, IB1), based on machine learning and principles of artificial intelligence, appear usable for these tasks.

On the nature of phonological cues in the acquisition of French gender categories: Evidence from instance-based learning models

A crucial element in acquiring a language involves partitioning its words into different grammatical categories. Although fundamentally the determination of grammatical gender in a language such as French is based on recognising certain distributional patterns, a noun's phonological ending has also been implicated as providing an important additional cue as to an unknown noun's gender value. Not only does the literature which explores this possibility rarely make precise the notion of “final sounds” but also those that are referred to appear not to be consonant with either phonological or psycholinguistic theory. This paper explores the suitability of the final syllable as a reliable cue. The issue is investigated through an instance-based machine learning model, in part because it more closely fits with psycholinguistic assumptions of gender processing. The results of the simulations show that not only does the final syllable prove a reliable indicator but that it is, in fact, more reliable than most other sequences. The paper, however, remains agnostic as to the exact role of analogising in either gender acquisition or gender processing. Tentative suggestions are made that gender attribution may require a dual mechanism account.

Robust Machine Learning Applied to Astronomical Data Sets. II. Quantifying Photometric Redshifts for Quasars Using Instance‐based Learning

We apply instance-based machine learning in the form of a k-nearest neighbor algorithm to the task of estimating photometric redshifts for 55,746 objects spectroscopically classified as quasars in the Fifth Data Release of the Sloan Digital Sky Survey. We compare the results obtained to those from an empirical color-redshift relation (CZR). In contrast to previously published results using CZRs, we find that the instance-based photometric redshifts are assigned with no regions of catastrophic failure. Remaining outliers are simply scattered about the ideal relation, in a similar manner to the pattern seen in the optical for normal galaxies at redshifts z < ~1. The instance-based algorithm is trained on a representative sample of the data and pseudo-blind-tested on the remaining unseen data. The variance between the photometric and spectroscopic redshifts is sigma^2 = 0.123 +/- 0.002 (compared to sigma^2 = 0.265 +/- 0.006 for the CZR), and 54.9 +/- 0.7%, 73.3 +/- 0.6%, and 80.7 +/- 0.3% of the objects are within delta z < 0.1, 0.2, and 0.3 respectively. We also match our sample to the Second Data Release of the Galaxy Evolution Explorer legacy data and the resulting 7,642 objects show a further improvement, giving a variance of sigma^2 = 0.054 +/- 0.005, and 70.8 +/- 1.2%, 85.8 +/- 1.0%, and 90.8 +/- 0.7% of objects within delta z < 0.1, 0.2, and 0.3. We show that the improvement is indeed due to the extra information provided by GALEX, by training on the same dataset using purely SDSS photometry, which has a variance of sigma^2 = 0.090 +/- 0.007. Each set of results represents a realistic standard for application to further datasets for which the spectra are representative.