Improving the Clustering Algorithms Automatic Generation Process with Cluster Quality Indexes

The two-dimensional transfer functions (TFs) designed based on intensity-gradient magnitude (IGM) histogram are effective tools for the visualization and exploration of 3D volume data. However, traditional design methods usually depend on multiple times of trial-and-error. We propose a novel method for the automatic generation of transfer functions by performing the affinity propagation (AP) clustering algorithm on the IGM histogram. Compared with previous clustering algorithms that were employed in volume visualization, the AP clustering algorithm has much faster convergence speed and can achieve more accurate clustering results. In order to obtain meaningful clustering results, we introduce two similarity measurements: IGM similarity and spatial similarity. These two similarity measurements can effectively bring the voxels of the same tissue together and differentiate the voxels of different tissues so that the generated TFs can assign different optical properties to different tissues. Before performing the clustering algorithm on the IGM histogram, we propose to remove noisy voxels based on the spatial information of voxels. Our method does not require users to input the number of clusters, and the classification and visualization process is automatic and efficient. Experiments on various datasets demonstrate the effectiveness of the proposed method.

Download Full-text

Cluster Quality Indexes for Symbolic Classification — An Examination

Studies in Classification, Data Analysis, and Knowledge Organization - Advances in Data Analysis ◽

10.1007/978-3-540-70981-7_4 ◽

2007 ◽

pp. 31-38

Author(s):

Andrzej Dudek

Keyword(s):

Quality Indexes ◽

Cluster Quality

Download Full-text

Analysis of Network Clustering Algorithms and Cluster Quality Metrics at Scale

PLoS ONE ◽

10.1371/journal.pone.0159161 ◽

2016 ◽

Vol 11 (7) ◽

pp. e0159161 ◽

Cited By ~ 57

Author(s):

Scott Emmons ◽

Stephen Kobourov ◽

Mike Gallant ◽

Katy Börner

Keyword(s):

Clustering Algorithms ◽

Quality Metrics ◽

Network Clustering ◽

Cluster Quality

Download Full-text

A Relative Performance of Dissimilarity Measures for Matching Relational Web Access Patterns Between User Sessions

Advances in Systems Analysis, Software Engineering, and High Performance Computing - Handbook of Research on Pattern Engineering System Development for Big Data Analytics ◽

10.4018/978-1-5225-3870-7.ch010 ◽

2018 ◽

pp. 153-176

Author(s):

Dilip Singh Sisodia

Keyword(s):

Clustering Algorithms ◽

Similarity Measures ◽

Comparative Performance ◽

Dissimilarity Measures ◽

User Session ◽

Relational Clustering ◽

Cluster Distance ◽

Web Access ◽

Access Patterns ◽

Cluster Quality

Customized web services are offered to users by grouping them according to their access patterns. Clustering techniques are very useful in grouping users and analyzing web access patterns. Clustering can be an object clustering performed on feature vectors or relational clustering performed on relational data. The relational clustering is preferred over object clustering for web users' sessions because of high dimensionality and sparsity of web users' data. However, relational clustering of web users depends on underlying dissimilarity measures used. Therefore, correct dissimilarity measure for matching relational web access patterns between user sessions is very important. In this chapter, the various dissimilarity measures used in relational clustering of web users' data are discussed. The concept of an augmented user session is also discussed to derive different augmented session dissimilarity measures. The discussed session dissimilarity measures are used with relational fuzzy clustering algorithms. The comparative performance binary session similarity and augmented session similarity measures are evaluated using intra-cluster and inter-cluster distance-based cluster quality ratio. The results suggested the augmented session dissimilarity measures in general, and intuitive augmented session (dis)similarity measure, in particular, performed better than the other measures.

Download Full-text

A system for deduction-based formal verification of workflow-oriented software models

International Journal of Applied Mathematics and Computer Science ◽

10.2478/amcs-2014-0069 ◽

2014 ◽

Vol 24 (4) ◽

pp. 941-956 ◽

Cited By ~ 10

Author(s):

Radosław Klimek

Keyword(s):

Formal Verification ◽

Temporal Logic ◽

Business Processes ◽

Deductive Reasoning ◽

Main Idea ◽

Automatic Generation ◽

Generation Process ◽

Software Models ◽

Logical Specifications ◽

Inexperienced User

Abstract The work concerns formal verification of workflow-oriented software models using the deductive approach. The formal correctness of a model’s behaviour is considered. Manually building logical specifications, which are regarded as a set of temporal logic formulas, seems to be a significant obstacle for an inexperienced user when applying the deductive approach. A system, along with its architecture, for deduction-based verification of workflow-oriented models is proposed. The process inference is based on the semantic tableaux method, which has some advantages when compared with traditional deduction strategies. The algorithm for automatic generation of logical specifications is proposed. The generation procedure is based on predefined workflow patterns for BPMN, which is a standard and dominant notation for the modeling of business processes. The main idea behind the approach is to consider patterns, defined in terms of temporal logic, as a kind of (logical) primitives which enable the transformation of models to temporal logic formulas constituting a logical specification. Automation of the generation process is crucial for bridging the gap between the intuitiveness of deductive reasoning and the difficulty of its practical application when logical specifications are built manually. This approach has gone some way towards supporting, hopefully enhancing, our understanding of deduction-based formal verification of workflow-oriented models.

Download Full-text

AUTOMATIC WINDING GENERATION USING MATRIX REPRESENTATION - ANFRACTUS TOOL 1.0

Acta Polytechnica ◽

10.14311/ap.2018.58.0037 ◽

2018 ◽

Vol 58 (1) ◽

pp. 37 ◽

Cited By ~ 3

Author(s):

Daoud Ouamara ◽

Frédéric Dubas ◽

Mohamed Nadjib Benallal ◽

Sid Ali Randi ◽

Christophe Espanet

Keyword(s):

Matrix Representation ◽

Automatic Generation ◽

Harmonic Spectrum ◽

Electrical Machines ◽

Generation Process ◽

Magnetomotive Force ◽

Research Software ◽

Input Parameters ◽

The Matrix ◽

Original Approach

This paper describes an original approach dealing with AC/DC winding design in electrical machines. A research software called “ANFRACTUS Tool 1.0”, allowing automatic generation of all windings in multi-phases electrical machines, has been developed using the matrix representation. Unlike existent methods, where the aim is to synthesize a winding with higher performances, the proposed method provides the opportunity to choose between all doable windings. The specificity of this approach is based on the fact that it take only the slots, phases and layers number as input parameters. The poles number is not requested to run the generation process. Windings generation by matrix representation may be applied for any number of slots, phases and layers. The software do not deal with the manner that coils are connected but just the emplacement of coils in each slot with its current sense. The waveform and the harmonic spectrum of the total magnetomotive force (MMF) are given as result.

Download Full-text

A self-learning finite element extraction system based on reinforcement learning

Artificial intelligence for engineering design analysis and manufacturing ◽

10.1017/s089006042100007x ◽

2021 ◽

pp. 1-29

Author(s):

Jie Pan ◽

Jingwei Huang ◽

Yunli Wang ◽

Gengdong Cheng ◽

Yong Zeng

Keyword(s):

Reinforcement Learning ◽

Mesh Generation ◽

Domain Knowledge ◽

Domain Boundary ◽

Automatic Generation ◽

Extraction System ◽

Generation Process ◽

High Quality ◽

Geometric Boundary ◽

Self Learning

Abstract Automatic generation of high-quality meshes is a base of CAD/CAE systems. The element extraction is a major mesh generation method for its capabilities to generate high-quality meshes around the domain boundary and to control local mesh densities. However, its widespread applications have been inhibited by the difficulties in generating satisfactory meshes in the interior of a domain or even in generating a complete mesh. The element extraction method's primary challenge is to define element extraction rules for achieving high-quality meshes in both the boundary and the interior of a geometric domain with complex shapes. This paper presents a self-learning element extraction system, FreeMesh-S, that can automatically acquire robust and high-quality element extraction rules. Two central components enable the FreeMesh-S: (1) three primitive structures of element extraction rules, which are constructed according to boundary patterns of any geometric boundary shapes; (2) a novel self-learning schema, which is used to automatically define and refine the relationships between the parameters included in the element extraction rules, by combining an Advantage Actor-Critic (A2C) reinforcement learning network and a Feedforward Neural Network (FNN). The A2C network learns the mesh generation process through random mesh element extraction actions using element quality as a reward signal and produces high-quality elements over time. The FNN takes the mesh generated from the A2C as samples to train itself for the fast generation of high-quality elements. FreeMesh-S is demonstrated by its application to two-dimensional quad mesh generation. The meshing performance of FreeMesh-S is compared with three existing popular approaches on ten pre-defined domain boundaries. The experimental results show that even with much less domain knowledge required to develop the algorithm, FreeMesh-S outperforms those three approaches in essential indices. FreeMesh-S significantly reduces the time and expertise needed to create high-quality mesh generation algorithms.

Download Full-text

Single-Cell Transcriptome Profiling Simulation Reveals the Impact of Sequencing Parameters and Algorithms on Clustering

Life ◽

10.3390/life11070716 ◽

2021 ◽

Vol 11 (7) ◽

pp. 716

Author(s):

Yunhe Liu ◽

Aoshen Wu ◽

Xueqing Peng ◽

Xiaona Liu ◽

Gang Liu ◽

...

Keyword(s):

Clustering Algorithms ◽

Simulated Data ◽

Ground Truth ◽

Real Data ◽

Transcriptome Profiling ◽

Actual Data ◽

Generation Process ◽

Data Generation ◽

The Matrix ◽

The Impact

Despite the scRNA-seq analytic algorithms developed, their performance for cell clustering cannot be quantified due to the unknown “true” clusters. Referencing the transcriptomic heterogeneity of cell clusters, a “true” mRNA number matrix of cell individuals was defined as ground truth. Based on the matrix and the actual data generation procedure, a simulation program (SSCRNA) for raw data was developed. Subsequently, the consistency between simulated data and real data was evaluated. Furthermore, the impact of sequencing depth and algorithms for analyses on cluster accuracy was quantified. As a result, the simulation result was highly consistent with that of the actual data. Among the clustering algorithms, the Gaussian normalization method was the more recommended. As for the clustering algorithms, the K-means clustering method was more stable than K-means plus Louvain clustering. In conclusion, the scRNA simulation algorithm developed restores the actual data generation process, discovers the impact of parameters on classification, compares the normalization/clustering algorithms, and provides novel insight into scRNA analyses.

Download Full-text

Auto Encoder and Deep Auto Encoders Based Deep Learning on Medical Data

International Journal of Recent Technology and Engineering - 2 ◽

10.35940/ijrte.d8205.118419 ◽

2019 ◽

Vol 8 (4) ◽

pp. 3832-3835

Keyword(s):

Neural Networks ◽

Feature Extraction ◽

Artificial Neural Networks ◽

Deep Learning ◽

Clustering Algorithms ◽

Medical Data ◽

Machine Learning Techniques ◽

Knowledge Generation ◽

Generation Process ◽

Artificial Neural

In rapid growth of medical informatics, patient data need to be organized and used for medical diagnosis and other uses such as disease prediction and drug discovery. There are many more traditional methods used for text based information such as K-NN, K-Means and other clustering algorithms, but image based medical data (or) signals based medical data is needed. So there is a need of new approaches for efficient classification and knowledge generation process. Artificial neural network based methods are mostly suited for deep learning, since there are many more approaches available in artificial neural networks. Deep learning and Machine learning techniques requires efficient pattern or feature extraction and pattern identification. Auto encoders and deep auto encoders works based on artificial neural networks and most suitable multimodal data feature extraction and identification. In this paper we have to show deep learning methods such as auto encoder and deep auto encoders for classifying multimodal medical data.

Download Full-text

CityJSON Building Generation from Airborne LiDAR 3D Point Clouds

ISPRS International Journal of Geo-Information ◽

10.3390/ijgi9090521 ◽

2020 ◽

Vol 9 (9) ◽

pp. 521

Author(s):

Gilles-Antoine Nys ◽

Florent Poux ◽

Roland Billen

Keyword(s):

Smart Cities ◽

Contextual Information ◽

Compliance Rate ◽

Point Clouds ◽

3D Models ◽

Automatic Generation ◽

Airborne Lidar ◽

Generation Process ◽

3D Point Clouds ◽

Unique Source

The relevant insights provided by 3D City models greatly improve Smart Cities and their management policies. In the urban built environment, buildings frequently represent the most studied and modeled features. CityJSON format proposes a lightweight and developer-friendly alternative to CityGML. This paper proposes an improvement to the usability of 3D models providing an automatic generation method in CityJSON, to ensure compactness, expressivity, and interoperability. In addition to a compliance rate in excess of 92% for geometry and topology, the generated model allows the handling of contextual information, such as metadata and refined levels of details (LoD), in a built-in manner. By breaking down the building-generation process, it creates consistent building objects from the unique source of Light Detection and Ranging (LiDAR) point clouds.

Download Full-text