METHOD FOR CLASSIFYING THE FRAGRANT QUALITY OF WOOD

The present invention provides a method for classifying the fragrant quality of wood comprises the steps of generating smell prints of at least one heated sample of the fragrant wood using an electronic nose (103); reducing the total amount of sensors of the electronic nose (103) to a significant number by using a principal component analysis and predicting the fragrant quality of the sample via a modified k-th nearest neighbour k-fold cross validation module, wherein the module comprises the steps of: computing the relationship between the sample data and previously tested sample data from which is stored in a database, selecting k previously tested samples which have the closest relationship with the sample from the database, characterized in that the step of predicting the fragrant quality of the sample further comprises the steps of classifying the fragrant quality of the sample using a voting principle among the k previously tested samples, and updating the database by adding the information of the sample to the database.