OPTIMASI METODE NAIVE BAYES MENGGUNAKAN METODE PARTICLE SWARM OPTIMIZATION (PSO) DALAM KLASIFIKASI PENDERITA PENYAKIT PARKINSON
Total View This Week0
Total View This Week0
Institusion
Universitas Sriwijaya
Author
R.M. ABURIZAL ARIF (STUDENT ID : 09021281419125)
Samsuryadi Samsuryadi (LECTURER ID : 0004027101)
Kanda Januar Miraswan (LECTURER ID : 0009019002)
Samsuryadi Samsuryadi (LECTURER ID : 0004027101)
Kanda Januar Miraswan (LECTURER ID : 0009019002)
Subject
R858-859.7 Computer applications to medicine. Medical informatics
Datestamp
2019-10-01 06:42:25
Abstract :
Naive Bayes is a classification method that is quite effective and widely used, but this method has disadvantages because of the nature of the independence of attributes that assume all attributes are the same so that calculations are carried out individually to determine the results of data classification. This has an effect on the value of accuracy produced. Therefore, optimization is needed to overcome the independent nature of the data. This study optimizes the Naive Bayes method with attribute weighting using Particle Swarm Optimization (PSO). The data used is the voice recording feature data of Parkinson's sufferers with a total number of 1040 data. The data is used because of the large amount of data which is suitable with the Naïve Bayes method. Testing is done by dividing it into 3 experimental configurations. The first experiment configuration was done by tuning the population that produced the best population number, namely 30 populations, the second experimental configuration tuning the number of generations with the most optimal number of generations are 40 generations, and the third experimental configuration comparing the optimization results using optimal parameters with classification before configuration optimization this third trial resulted in an accuracy value of Parkinson's sufferers data classification of 68.08%. The increase in average classification accuracy reaches 6.2% of the value of accuracy before optimization. The maximum accuracy value when the Naive Bayes method is optimized with PSO is reached 73.08%. The weighting of attributes performed by PSO succeeded in increasing the accuracy of the Naïve Bayes method in classifying data of Parkinson's sufferers.
Naive Bayes is a classification method that is quite effective and widely used, but this method has disadvantages because of the nature of the independence of attributes that assume all attributes are the same so that calculations are carried out individually to determine the results of data classification. This has an effect on the value of accuracy produced. Therefore, optimization is needed to overcome the independent nature of the data. This study optimizes the Naive Bayes method with attribute weighting using Particle Swarm Optimization (PSO). The data used is the voice recording feature data of Parkinson's sufferers with a total number of 1040 data. The data is used because of the large amount of data which is suitable with the Naïve Bayes method. Testing is done by dividing it into 3 experimental configurations. The first experiment configuration was done by tuning the population that produced the best population number, namely 30 populations, the second experimental configuration tuning the number of generations with the most optimal number of generations are 40 generations, and the third experimental configuration comparing the optimization results using optimal parameters with classification before configuration optimization this third trial resulted in an accuracy value of Parkinson's sufferers data classification of 68.08%. The increase in average classification accuracy reaches 6.2% of the value of accuracy before optimization. The maximum accuracy value when the Naive Bayes method is optimized with PSO is reached 73.08%. The weighting of attributes performed by PSO succeeded in increasing the accuracy of the Naïve Bayes method in classifying data of Parkinson's sufferers.
File :
RAMA_55201_09021281419125_0004027101_0009019002_01_front_ref.pdf
RAMA_55201_09021281419125_0004027101_0009019002_02.pdf
RAMA_55201_09021281419125_0004027101_0009019002_03.pdf
RAMA_55201_09021281419125_0004027101_0009019002_04.pdf
RAMA_55201_09021281419125_0004027101_0009019002_05.pdf
RAMA_55201_09021281419125_0004027101_0009019002_06.pdf
RAMA_55201_09021281419125_0004027101_0009019002_07_ref.pdf
RAMA_55201_09021281419125_0004027101_0009019002_08_lamp.pdf
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
