Structure Learning of Bayesian Networks Using Elephant Swarm Water Search Algorithm

2020 ◽  
Vol 11 (2) ◽  
pp. 19-30
Author(s):  
Shahab Wahhab Kareem ◽  
Mehmet Cudi Okur
Keyword(s):  
Bayesian Networks ◽  
Structure Learning ◽  
Search Algorithm ◽  
Confusion Matrix ◽  
Score Function ◽  
Analytical Models ◽  
Data Sets ◽  
Novel Approach ◽  
Networks Structure ◽  
Structure Of Knowledge

Bayesian networks are useful analytical models for designing the structure of knowledge in machine learning. Bayesian networks can represent probabilistic dependency relationships among the variables. One strategy of Bayesian Networks structure learning is the score and search technique. The authors present the Elephant Swarm Water Search Algorithm (ESWSA) as a novel approach to Bayesian network structure learning. In the algorithm; Deleting, Reversing, Inserting, and Moving are used to make the ESWSA for reaching the optimal structure solution. Mainly, water search strategy of elephants during drought periods is used in the ESWSA algorithm. The proposed method is compared with simulated annealing and greedy search using BDe score function. The authors have also investigated the confusion matrix performances of these techniques utilizing various benchmark data sets. As presented by the results of the evaluations, the proposed algorithm has better performance than the other algorithms and produces better scores and accuracy values.

scholarly journals Falcon Optimization Algorithm for Bayesian Networks Structure Learning

2021 ◽  
Vol 22 (4) ◽  
Author(s):  
Shahab Wahhab Kareem ◽  
Mehmet Cudi Okur
Keyword(s):  
Bayesian Networks ◽  
Bayesian Network ◽  
Optimization Algorithm ◽  
Structure Learning ◽  
Confusion Matrix ◽  
Optimal Solution ◽  
Score Function ◽  
Scientific Models ◽  
Data Sets ◽  
Networks Structure

In machine-learning, one of the useful scientific models for producing the structure of knowledge is Bayesian network, which can draw probabilistic dependency relationships between variables. The score and search is a method used for learning the structure of a Bayesian network. The authors apply the Falcon Optimization Algorithm (FOA) as a new approach to learning the structure of Bayesian networks. This paper uses the Reversing, Deleting, Moving and Inserting operations to adopt the FOA for approaching the optimal solution of Bayesian network structure. Essentially, the falcon prey search strategy is used in the FOA algorithm. The result of the proposed technique is compared with Pigeon Inspired optimization, Greedy Search, and Simulated Annealing using the BDeu score function. The authors have also examined the performances of the confusion matrix of these techniques utilizing several benchmark data sets. As shown by the evaluations, the proposed method has more reliable performance than the other algorithms including producing better scores and accuracy values.

Evaluation of Bayesian Network Structure Learning Using Elephant Swarm Water Search Algorithm

2020 ◽  
pp. 139-159
Author(s):  
Shahab Wahhab Kareem ◽  
Mehmet Cudi Okur
Keyword(s):  
Simulated Annealing ◽  
Bayesian Network ◽  
Network Structure ◽  
Structure Learning ◽  
Search Algorithm ◽  
Confusion Matrix ◽  
Data Sets ◽  
Greedy Search ◽  
Bayesian Network Structure ◽  
Bayesian Network Structure Learning

Bayesian networks are useful analytical models for designing the structure of knowledge in machine learning which can represent probabilistic dependency relationships among the variables. The authors present the Elephant Swarm Water Search Algorithm (ESWSA) for Bayesian network structure learning. In the algorithm; Deleting, Reversing, Inserting, and Moving are used to make the ESWSA for reaching the optimal structure solution. Mainly, water search strategy of elephants during drought periods is used in the ESWSA algorithm. The proposed method is compared with Pigeon Inspired Optimization, Simulated Annealing, Greedy Search, Hybrid Bee with Simulated Annealing, and Hybrid Bee with Greedy Search using BDeu score function as a metric for all algorithms. They investigated the confusion matrix performances of these techniques utilizing various benchmark data sets. As presented by the results of evaluations, the proposed algorithm achieves better performance than the other algorithms and produces better scores as well as the better values.

scholarly journals Detection of natural structures and classification of HCI-HPR data using robust forward search algorithm

2016 ◽  
Vol 9 (1) ◽  
pp. 23-41 ◽  
Author(s):  
Fatima Isiaka ◽  
Kassim S Mwitondi ◽  
Adamu M Ibrahim
Keyword(s):  
Search Algorithm ◽  
Original Data ◽  
Data Sets ◽  
Forward Search ◽  
Distribution Method ◽  
Mahalanobis Distances ◽  
Content Type ◽  
Movement Data ◽  
Novel Approach ◽  
Fitting In

Purpose – The purpose of this paper is to proposes a forward search algorithm for detecting and identifying natural structures arising in human-computer interaction (HCI) and human physiological response (HPR) data. Design/methodology/approach – The paper portrays aspects that are essential to modelling and precision in detection. The methods involves developed algorithm for detecting outliers in data to recognise natural patterns in incessant data such as HCI-HPR data. The detected categorical data are simultaneously labelled based on the data reliance on parametric rules to predictive models used in classification algorithms. Data were also simulated based on multivariate normal distribution method and used to compare and validate the original data. Findings – Results shows that the forward search method provides robust features that are capable of repelling over-fitting in physiological and eye movement data. Research limitations/implications – One of the limitations of the robust forward search algorithm is that when the number of digits for residuals value is more than the expected size for stack flow, it normally yields an error caution; to counter this, the data sets are normally standardized by taking the logarithmic function of the model before running the algorithm. Practical implications – The authors conducted some of the experiments at individual residence which may affect environmental constraints. Originality/value – The novel approach to this method is the detection of outliers for data sets based on the Mahalanobis distances on HCI and HPR. And can also involve a large size of data with p possible parameters. The improvement made to the algorithm is application of more graphical display and rendering of the residual plot.

A novel approach to fully representing the diversity in conditional dependencies for learning Bayesian network classifier

2021 ◽  
Vol 25 (1) ◽  
pp. 35-55
Author(s):  
Limin Wang ◽  
Peng Chen ◽  
Shenglei Chen ◽  
Minghui Sun
Keyword(s):  
Bayesian Network ◽  
Conditional Independence ◽  
Structure Learning ◽  
Classification Performance ◽  
Data Sets ◽  
Independence Assumption ◽  
Bayesian Network Classifiers ◽  
Novel Approach ◽  
Conditional Independence Assumption ◽  
Dependence Criterion

Bayesian network classifiers (BNCs) have proved their effectiveness and efficiency in the supervised learning framework. Numerous variations of conditional independence assumption have been proposed to address the issue of NP-hard structure learning of BNC. However, researchers focus on identifying conditional dependence rather than conditional independence, and information-theoretic criteria cannot identify the diversity in conditional (in)dependencies for different instances. In this paper, the maximum correlation criterion and minimum dependence criterion are introduced to sort attributes and identify conditional independencies, respectively. The heuristic search strategy is applied to find possible global solution for achieving the trade-off between significant dependency relationships and independence assumption. Our extensive experimental evaluation on widely used benchmark data sets reveals that the proposed algorithm achieves competitive classification performance compared to state-of-the-art single model learners (e.g., TAN, KDB, KNN and SVM) and ensemble learners (e.g., ATAN and AODE).

scholarly journals Improved Local Search with Momentum for Bayesian Networks Structure Learning

Entropy ◽  
2021 ◽  
Vol 23 (6) ◽  
pp. 750
Author(s):  
Xiaohan Liu ◽  
Xiaoguang Gao ◽  
Zidong Wang ◽  
Xinxin Ru
Keyword(s):  
Local Search ◽  
Complex Networks ◽  
Bayesian Networks ◽  
Structure Learning ◽  
Hybrid Methods ◽  
State Of The Art ◽  
The State ◽  
Experimental Results ◽  
Local Optimum ◽  
Networks Structure

Bayesian Networks structure learning (BNSL) is a troublesome problem that aims to search for an optimal structure. An exact search tends to sacrifice a significant amount of time and memory to promote accuracy, while the local search can tackle complex networks with thousands of variables but commonly gets stuck in a local optimum. In this paper, two novel and practical operators and a derived operator are proposed to perturb structures and maintain the acyclicity. Then, we design a framework, incorporating an influential perturbation factor integrated by three proposed operators, to escape current local optimal and improve the dilemma that outcomes trap in local optimal. The experimental results illustrate that our algorithm can output competitive results compared with the state-of-the-art constraint-based method in most cases. Meanwhile, our algorithm reaches an equivalent or better solution found by the state-of-the-art exact search and hybrid methods.

scholarly journals Prediksi Produktivitas Tanaman Padi di Kabupaten Karawang Menggunakan Bayesian Networks

2017 ◽  
Vol 9 (4) ◽  
Author(s):  
Betha Nurina Sari ◽  
Hendi Permana ◽  
Kardo Trihandoko ◽  
Asep Jamaludin ◽  
Yuyun Umaidah
Keyword(s):  
Bayesian Networks ◽  
Learning Process ◽  
Structure Learning ◽  
Confusion Matrix ◽  
Numerical Data ◽  
Training Data ◽  
Nominal Data ◽  
Productivity Level ◽  
Rice Productivity ◽  
Data Discretization

This research is aimed to build a model for predicting rice productivity level in Karawang district. The prediction using Bayesian Networks allowed three stages, pre-processing of data, implementation and evaluation stages. Pre-processing is transformation of numerical data into nominal data by using two scenarios, using threshold mean and discretization. Implementation stage is to apply Bayesian Networks algorithm, that is through structure learning process and parameter learning. The learning process of structures and parameters on bayesian networks using CaMML 1.41 software. Evaluation of Bayesian Networks performance in predicting rice productivity with confusion matrix, ie calculating prediction accuracy and log loss. The experiment results show the satisfactory results, the accuracy above 90%. The best model generated from pre-processing using the data discretization and 5-year training and 1-year testing data. This explain that the selection techniques of pre-processing and the technique of dividing the training data and testing the data affect the results of the performance evaluation of the structure of Bayesian Networks.

Sign in / Sign up

Export Citation Format

Share Document