Automotive Vehicles Quality Prediction based on Features Customization and Differentiators using Artificial Neural Network in Comparison with Digraph Approach.

Aim: The aim of the proposed work is to predict the performance of an Artificial Neural Network (ANN) algorithm in detection of vehicle quality performance based on Features customization and differentiators by comparing it with the Digraph algorithm. Materials and Methods:The proposed ANN is trained and tested with a "Novel Car Evaluation Database" created by Marko Bohanec. With correct quality 826 samples and inaccurate quality 826 samples in two groups with a total sample size of 1652. Training data [75% of dataset] and testing data set [25% of data set] are separated from the obtained samples. The samples are calculated using G power analysis using clincalc, which includes two groups: alpha (0.05), power (80%). There is a statistically significant difference between the groups with p=0.04 Results: The proposed ANN algorithm provided better results in predicting the quality of novel car evaluation compared to the Digraph approach. Conclusion: For the given data set, from the results it is found that the ANN algorithm is significantly more suitable for recognition of better car quality. [ABSTRACT FROM AUTHOR]