A one-class classification decision tree based on kernel density estimation

Itani, Sarah; Lecron, Fabian; Fortemps, Philippe

doi:10.1016/j.asoc.2020.106250

Cited by 29 publications

(11 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The problem of pocket retrieval thus appears as an instance of a one-class discrimination problem [28]. One-class discrimination is a learning task that typically arises in outlier (anomaly) detection or, more generally, in binary discrimination data mining problems where obtaining examples of one class can be too expensive or daunting, or where examples of one class are largely under represented (data imbalance) [29,30]. Different approaches are used in the literature to solve one-class or data imbalance problems.…”

Section: Discussionmentioning

confidence: 99%

SHREC 2022: Protein-ligand binding site recognition

Gagliardi¹,

Raffo²,

Fugacci³

et al. 2022

Preprint

View full text Add to dashboard Cite

This paper presents the methods that have participated in the SHREC 2022 contest on protein-ligand binding site recognition. The prediction of protein-ligand binding regions is an active research domain in computational biophysics and structural biology and plays a relevant role for molecular docking and drug design. The goal of the contest is to assess the effectiveness of computational methods in recognizing ligand binding sites in a protein based on its geometrical structure. Performances of the segmentation algorithms are analyzed according to two evaluation scores describing the capacity of a putative pocket to contact a ligand and to pinpoint the correct binding region. Despite some methods perform remarkably, we show that simple non-machinelearning approaches remain very competitive against data-driven algorithms. In general, the task of pocket detection remains a challenging learning problem which suffers of intrinsic difficulties due to the lack of negative examples (data imbalance problem).

show abstract

Section: Discussionmentioning

confidence: 99%

SHREC 2022: Protein-ligand binding site recognition

Gagliardi¹,

Raffo²,

Fugacci³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…In Itani et al., 22 by combining kernel density estimation with decision tree, the data density distribution of kernel density estimation is used as the division criterion of decision number, and the boundary width of each feature is determined by forming the final decision tree structure, thus forming the super rectangle discriminant boundary. This method determines the coefficient multiplied by

{N_{total}}^{-{1/5}}

according to the standard deviation and interquartile range of the sample.…”

Section: Fast Decision Algorithm Designmentioning

confidence: 99%

A clustering hyperrectangle model based on kernel density estimation for product quality detection

Lü

Xie

Tang

2022

Quality & Reliability Eng

View full text Add to dashboard Cite

With the rapid development of sensor technology, a huge amount of data is generated in the industrial manufacturing process, which poses new challenges for product quality detection. Meanwhile, the process data often have high‐dimensional characteristics of nonlinearity and strong correlation, so traditional methods are not suitable for quick quality detection. In this paper, to further enhance the speed and accuracy of quality detection, a clustering hyperrectangle model based on a kernel density estimation is proposed. First, based on the kernel density estimation, the distribution characteristics of data, which can directly provide optimal clustering parameter for the k‐means method, can be obtained. Next, to form the detection boundary of every hyperrectangle, the mean vector formed by clustering is used as the basis for determining the center of the hyperrectangle. Then, the distance between every sample point and the center of hyperrectangle is calculated for building the final detection condition. Furthermore, the joint determination based on the detection condition is used for effectively guaranteeing the detection accuracy. Finally, based on the simulation and experimental results, compared with the traditional methods, the proposed method can determine the detection condition in a more refined way, and has the merits of shorter test time and higher accuracy.

show abstract

“…Different types of classification methods, such as artificial neural networks [35,36], SVMs [37], random forest [17], decision trees [38] and k-nearest neighbors [39], can be used to classify the CTI functioning/failed state from monitoring signals. In this work, SVM classifiers have been adopted, since they offer high classification performance, low computational cost, and ability to handle imbalanced datasets by using its Cost-Sensitive adjustment [26,40].…”

Section: Classifiermentioning

confidence: 99%

Identification of Critical Components in the Complex Technical Infrastructure of the Large Hadron Collider Using Relief Feature Ranking and Support Vector Machines

et al. 2021

View full text Add to dashboard Cite

This work proposes a data-driven methodology for identifying critical components in Complex Technical Infrastructures (CTIs), for which the functional logic and/or the system structure functions are not known due the CTI’s complexity and evolving nature. The methodology uses large amounts of CTI monitoring data acquired over long periods of time and under different operating conditions. The critical components are identified as those for which the condition monitoring signals permit the optimal classification of the CTI functioning or failed state. The methodology includes two stages: in the first stage, a feature selection filter method based on the Relief technique is used to rank the monitoring signals according to their importance with respect to the CTI functioning or failed state; the second stage identifies the subset of signals among those highlighted by the Relief technique that are most informative with respect to the CTI state. This identification is performed on the basis of evaluating the performance of a Cost-Sensitive Support Vector Machine (CS-SVM) classifier trained with several subsets of the candidate signals. The capabilities of the methodology proposed are assessed through its application to different benchmarks of highly imbalanced datasets, showing performances that are competitive to those obtained by other methods presented in the literature. The methodology is finally applied to the monitoring signals of the Large Hadron Collider (LHC) of the European Organization for Nuclear Research (CERN), a CTI for experiments of physics; the criticality of the identified components has been confirmed by CERN experts.

show abstract

A one-class classification decision tree based on kernel density estimation

Cited by 29 publications

References 57 publications

SHREC 2022: Protein-ligand binding site recognition

SHREC 2022: Protein-ligand binding site recognition

A clustering hyperrectangle model based on kernel density estimation for product quality detection

Identification of Critical Components in the Complex Technical Infrastructure of the Large Hadron Collider Using Relief Feature Ranking and Support Vector Machines

Contact Info

Product

Resources

About