Clasificación automática de plantas monocotiledóneas y dicotiledóneas usando minería de datos

This research project, it describes based on a web application development within the two data mining techniques comparison, such as: logistic regression and SVM (support vector machine). For this case study, it was performed the field research, where was got the images for the database creation with 353 records. Into web application development is collected data, such as: area, perimeter, centroid and type (monocotyledonous and dicotyledonous), these data are used in the two aforementioned algorithms training and learning process; since it is useful for automatic classification. For the prototype development was used the image segmentation, morphological operations for the leaf recognition and subsequently, it is extracted the same attributes, said attributes are saved in a cvs, what used two models, through the functions model = LogisticRegression () and clf = SVC(kernel="rbf").fit(X_train, y_train), to get as a result, the plant classification, this can be (monocotyledonous and dicotyledonous). At the end, it gives a classification validation accuracy in the monocotyledonous and dicotyledonous plant with logistic 97.75% regression and in 73.03%SVM, whose shows that the data mining technique with the least error is logistic regression.

• 第一著者: Cayambe Cajo, Fabian Rolando (author)
• フォーマット: bachelorThesis
• 言語: spa
• 出版事項: 2022
• 主題: REGRESIÓN LOGÍSTICA; SMV; MINERÍA DE DATOS; MONOCOTILEDÓNEA Y DICOTILEDÓNEA; SISTEMAS COMPUTACIONALES
• オンライン･アクセス: http://repositorio.utc.edu.ec/handle/27000/9176