| dc.description.abstract | In agriculture, the use of spectroscopy on quality has long been novel for its effectiveness.
This dissertation compares different dimensionality reduction onto the dataset of
spectroscopy of mangos. From 58 samples of mangos, Agus Arip Munawar, Munawar,
Hayati, & Fakhruddin (2020) [1] provided enhanced dataset for their inner attributes
collected from Near Infrared Spectroscopy (NIRS) tool. Despite Principal Component
Analysis (PCA) and Partial Least Squares (PLSR) in the main paper, this thesis chooses to
introduce PCA. Later, Total Acidity’s mathematical indexes were measured and compared
to the available ones in the main paper. Other methods to find out the best features include
the novel Forward Feature Selection and Backward Elimination construction. Comparison
of them on Excel and Python will be made to justify which model provides better results
by the measurement of accuracy. This approach would open doors for new algorithms and
models to increase the prediction of mangoes spectroscopic data in the future. | en_US |
