Confusion Matrix Disagreement for Multiple Classifiers

Freitas, Cinthia O. A.; Carvalho, João; Oliveira, J.J. de; Aires, Simone B. K.; Sabourin, Robert

doi:10.1007/978-3-540-76725-1_41

Cited by 16 publications

(10 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Therefore, lacking a biologically-plausible similarity measure, we use the distance between the confusion matrices of the human observers and the learned model (cf. [16]). The model by Weijer et al has a distance of 0.73 whereas our model has a distance of 0.57.…”

Section: B Resultsmentioning

confidence: 93%

Web-Based Learning of Naturalized Color Models for Human-Machine Interaction

Schauerte

Fink

2010

2010 International Conference on Digital Image Computing: Techniques and Applications

View full text Add to dashboard Cite

In recent years, natural verbal and non-verbal human-robot interaction has attracted an increasing interest. Therefore, models for robustly detecting and describing visual attributes of objects such as, e.g., colors are of great importance. However, in order to learn robust models of visual attributes, large data sets are required. Based on the idea to overcome the shortage of annotated training data by acquiring images from the Internet, we propose a method for robustly learning natural color models. Its novel aspects with respect to prior art are: firstly, a randomized HSL transformation that reflects the slight variations and noise of colors observed in real-world imaging sensors; secondly, a probabilistic ranking and selection of the training samples, which removes a considerable amount of outliers from the training data. These two techniques allow us to estimate robust color models that better resemble the variances seen in realworld images. The advantages of the proposed method over the current state-of-the-art technique using the training data without proper transformation and selection are confirmed in experimental evaluations. In combination, for models learned with pLSA-bg and HSL, the proposed techniques reduce the amount of mislabeled objects by 19.87 % on the well-known E-Bay data set.

show abstract

Section: B Resultsmentioning

confidence: 93%

Web-Based Learning of Naturalized Color Models for Human-Machine Interaction

Schauerte

Fink

2010

2010 International Conference on Digital Image Computing: Techniques and Applications

View full text Add to dashboard Cite

show abstract

“…The configuration matrix is a table consisting of the number of rows of test data that are predicted to be true and false by the classification model used. A confusion matrix table is needed to select the best performance from a classification model [19]. This method is often used with multiple classifiers or more than two classes.…”

Section: E Confusion Matrixmentioning

confidence: 99%

A Comparison of Naive Bayes Methods, Logistic Regression and KNN for Predicting Healing of Covid-19 Patients in Indonesia

Romadhon

Kurniawan

2021

2021 3rd East Indonesia Conference on Computer and Information Technology (EIConCIT)

View full text Add to dashboard Cite

Since it was declared a global pandemic by the World Health Organization (WHO), the number of cases of Covid-19 patients who died has continued to increase. One of the countries with the highest death rate in the world is Indonesia. On Saturday, April 4, 2020, Indonesia reached the highest death rate for Covid-19 patients, around 9.11%. This number must be suppressed so that there are no more victims. For this reason, it is necessary to know actually the factors that can reduce the risk of death and predict the chance of curing Covid-19 patients. In data mining, there are several methods that can be used to predict a patient's recovery rate by considering several variables. The variables used in this study were age and gender. Naive Bayes Method, logistic regression, and K-Nearest Neighbor (KNN) are the methods to be chosen in this study to analyze their most accurate performance. The result shows that KNN has the highest accuracy, which is 0.750 compared to logistic regression which has a value of 0.703 as well as Naive Bayes which has the same value. Meanwhile, the level of precision of the three models shows that KNN also has the highest value, namely 0.750 than logistic regression and Naive Bayes which have the same value, namely 0.700. The recall value of the three also shows that the kNN remains the highest with 0.750 compared to the two comparison models which have the same value, namely 0.708.

show abstract

“…The accuracy of a classifier can be easily obtained from its corresponding confusion matrix. In their work (Freitas, de Carvalho, & Oliveira, 2007), the authors defined a simple measure on confusion matrix, i.e. the global performance index, for evaluating performances of ensemble classifiers.…”

Section: Some Measures Defined On Confusion Matrixmentioning

confidence: 99%