Applications of Naïve Bayesian Classifiers to linearly separable and inseparable data
Description (Abstract)
Classification is a supervised technique that trains a machine learning model using data objects with known classes and estimates classes of data objects without known classes. It is a powerful technique that helps decision making in various fields, for example, biology and finance. One of challenges associated with classification is the analysis of linearly inseparable data. As the number of features increases, nonlinearity of the data tends to be increased further. Naïve Bayesian (NB) classification method was developed based on probability theories. The method is highly scalable because it assumes that features are conditionally independent. This study shows that the NB classifiers returned high prediction accuracies on both linearly separable and linearly inseparable data.
Applications of Naïve Bayesian Classifiers to linearly separable and inseparable data
H239, 2nd floor, Barbara H. Hagan Center for Nursing
Classification is a supervised technique that trains a machine learning model using data objects with known classes and estimates classes of data objects without known classes. It is a powerful technique that helps decision making in various fields, for example, biology and finance. One of challenges associated with classification is the analysis of linearly inseparable data. As the number of features increases, nonlinearity of the data tends to be increased further. Naïve Bayesian (NB) classification method was developed based on probability theories. The method is highly scalable because it assumes that features are conditionally independent. This study shows that the NB classifiers returned high prediction accuracies on both linearly separable and linearly inseparable data.
