Using Supervised Machine Learning to Predict the Status of Road Signs

There is no data collected and saved about road signs in Sweden the status for these unknown. Furthermore, of sign colors, quality sign, type retroreflection material, age are Therefore, it difficult to know (approved or not) any without performing a costly inspection. The aim this study predict mounted on Swedish roads by using supervised machine learning. This investigates effect principal component analysis (PCA) scaling accuracy prediction. were prepared before then scaled two methods which normalization standardization. three algorithms that tested Random Forest, Artificial Neural Network (ANN), Support Vector Machines (SVM). They invoked signs. exhibited overall high predicting (98%), precision recall F1 scores (98%). forest showed best performance with 4 PC components normalized highest 98%. Using PCA different impacts techniques. In case ANN, invoking improves accuracy, while SVM decreases when used. On other hand, has random model invoked. standardization also investigated study, noticed increases prediction all models (ANN, Forest). better achieved compared normalization.