Comparative Analysis of Algorithms in Supervised Classification: A Case study of Bank Notes Dataset

Ghazvini, Anahita; Awwalu, Jamilu; Bakar, Azuraliza Abu

doi:10.14445/22312803/ijctt-v17p109

Cited by 15 publications

(7 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As the ten datasets are well known and are frequently used for evaluation of classification methods (Ghazvini et al 2014;Duch et al 2012;Tao et al 2004), we compare the performance of the δ-machine against the previous results. In general, there was not a single classification method that had significantly best results for all these datasets in the past, because each classification method relies on assumptions, estimators, or approximations.…”

Section: Resultsmentioning

confidence: 99%

“…Any classification method can achieve the best performance if there is an ideal dataset that fulfills these (Duin et al 2010). The three studies carried out by Duch et al (2012), Tao et al (2004), and Ghazvini et al (2014) had seven, two, and one out of ten datasets overlapping with our ten datasets, respectively. From these studies, we conclude that in Ghazvini et al (2014), and Tao et al (2004) five out of the ten datasets, SVM(RBF) had achieved the lowest MR. SVM with a linear kernel, multilayer perceptron networks (Hornik et al 1989), the iterative axis-parallel rectangle algorithm (Dietterich et al 1997) and kernel-based multiple-instance learning model (Tao et al 2004) had one best result each.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

The δ-Machine: Classification Based on Distances Towards Prototypes

2019

View full text Add to dashboard Cite

We introduce the δ-machine, a statistical learning tool for classification based on (dis)similarities between profiles of the observations to profiles of a representation set consisting of prototypes. In this article, we discuss the properties of the δ-machine, propose an automatic decision rule for deciding on the number of clusters for the K-means method on the predictive perspective, and derive variable importance measures and partial dependence plots for the machine. We performed five simulation studies to investigate the properties of the δ-machine. The first three simulation studies were conducted to investigate selection of prototypes, different (dis)similarity functions, and the definition of representation set. Results indicate that we best use the Lasso to select prototypes, that the Euclidean distance is a good dissimilarity function, and that finding a small representation set of prototypes gives sparse but competitive results. The remaining two simulation studies investigated the performance of the δ-machine with imbalanced classes and with unequal covariance matrices for the two classes. The results obtained show that the δ-machine is robust to class imbalances, and that the four (dis)similarity functions had the same performance regardless of the covariance matrices. We also showed the classification performance of the δ-machine compared with three other classification methods on ten real datasets from UCI database, and discuss two empirical examples in detail.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

The δ-Machine: Classification Based on Distances Towards Prototypes

2019

View full text Add to dashboard Cite

show abstract

“…Some paper has the same goal to measure accuracy in each method used for classification, especially classification for banknotes. These methods are Naïve Bayes and Multilayer Perceptron [6]; Probabilistic neural network (PNN), Multi-layer Perceptron (MLP), Radial Basis Function (RBF), Decision Tree (DT), and Naïve Base [3]; Backpropagation Neural Network (BPN) and Support Vector Machine (SVM) [1].…”

Section: Literature Reviewmentioning

confidence: 99%

“…The level accuracy of each method is used differently. In the Naïve Bayes method and Multilayer Perceptron method, the accuracy rate is 95% and 97%, so the Multilayer Perceptron method outperforms the Naïve Bayes method [6]. Whereas in the comparison of the Probabilistic neural network (PNN), Multi-layer Perceptron (MLP), Radial Basis Function (RBF), Decision Tree (DT), and Naïve Bayes method, the best accuracy of the method is Decision Tree (DT) method.…”

Section: Literature Reviewmentioning

confidence: 99%

“…The techniques in machine learning can be used to overcome various problems like classification, clustering, association rule mining, and prediction. The accuracy in an election the method becomes the primary key to overcome the problem is happening [6]. The type of machine learning used by authors is supervised Learning that aims to overcome the classification problem.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Influence Distribution Training Data on Performance Supervised Machine Learning Algorithms

Suban

Emanuel

2020

2020 3rd International Seminar on Research of Information Technology and Intelligent Systems (ISRITI)

View full text Add to dashboard Cite

Almost all fields of life need Banknote. Even particular fields of life require banknotes in large quantities such as banks, transportation companies, and casinos. Therefore Banknotes are an essential component in carrying out all activities every day, especially those related to finance. Through technological advancements such as scanners and copy machine, it can provide the opportunity for anyone to commit a crime. The crime is like a counterfeit banknote. Many people still find it difficult to distinguish between a genuine banknote and counterfeit Banknote, that is because counterfeit Banknote produced have a high degree of resemblance to the genuine Banknote. Based on that background, authors want to do a classification process to distinguish between genuine Banknote and counterfeit Banknote. The classification process use methods Supervised Learning and compares the level of accuracy based on the distribution of training data. The methods of supervised Learning used are Support Vector Machine (SVM), K-Nearest Neighbor (K-NN), and Naïve Bayes. K-NN method is a method that has the highest specificity, sensitivity, and accuracy of the three methods used by the authors both in the training data of 30%, 50%, and 80%. Where in the training data 30% and 50% value specificity: 0.99, sensitivity: 1.00, accuracy: 0.99. While the 80% training data value specificity: 1.00, sensitivity: 1.00, accuracy: 1.00. This means that the distribution of training data influences the performance of the Supervised Machine Learning algorithm. In the KNN method, the greater the training data, the better the accuracy.

show abstract

Identification of Genuine from Fake Banknotes Using an Enhanced Machine Learning Approach

Lyaqini

Nachaoui

2021

Springer Proceedings in Mathematics &Amp; Statistics

View full text Add to dashboard Cite

Comparative Analysis of Algorithms in Supervised Classification: A Case study of Bank Notes Dataset

Cited by 15 publications

References 0 publications

The δ-Machine: Classification Based on Distances Towards Prototypes

The δ-Machine: Classification Based on Distances Towards Prototypes

Influence Distribution Training Data on Performance Supervised Machine Learning Algorithms

Identification of Genuine from Fake Banknotes Using an Enhanced Machine Learning Approach

Contact Info

Product

Resources

About