Nama : Faris Abdi El Hakim
NRP: 122180006
TUGAS (HOMEWORK)
1. Pilihlah sedikitnya 5 datasets pada UCI Machine Learning reporitories
http://archive.ics.uci.edu/ml/
Dataset name Link
Iris http://archive.ics.uci.edu/ml/datasets/Iris
Wine http://archive.ics.uci.edu/ml/datasets/Wine
Blood http://archive.ics.uci.edu/ml/datasets/Blood+Transfusion+Service+Center
Transfusion
Service Center
Heart failure http://archive.ics.uci.edu/ml/datasets/Heart+failure+clinical+records
clinical records
Data Set
Haberman http://archive.ics.uci.edu/ml/datasets/Haberman%27s+Survival
Survival
2. Jalankan SVM pada semua dataset menggunakan WEKA Knowledge Flow dan catatlah hasil
eksperimen dalam sebuah tabel
No Dataset Hasil
1 Iris Hasil Akurasi : 82.6667%
Kappa statistic:0.74
MAE:0.1156
RMSE:0.3399
Weighted Avg:
TP Rate FP Rate Precision Recall F-Measure MCC ROC Area PRC Area
0.827 0.087 0.852 0.827 0.831 0.753 0.870 0.758
=== Confusion Matrix ===
a b c <-- classified as
41 1 8 | a = Iris-setosa
0 38 12 | b = Iris-versicolor
0 5 45 | c = Iris-virginica
2 Wine Hasil Akurasi : 81.9209 %
Kappa statistic: 0.7216
MAE: 0.2699
RMSE: 0.349
Weighted Avg:
TP Rate FP Rate Precision Recall F-Measure MCC ROC Area PRC Area
0.819 0.107 0.830 0.819 0.817 0.726 0.887 0.761
=== Confusion Matrix ===
a b c <-- classified as
51 6 1 | a = 1
7 62 2 | b = 2
2 14 32 | c = 3
3 Blood Hasil Akurasi : 75.4011 %
Transfusion Kappa statistic: 0.1171
Service MAE: 0.246
Center RMSE: 0.496
Weighted Avg.
TP Rate FP Rate Precision Recall F-Measure MCC ROC Area PRC Area
0.754 0.664 0.701 0.754 0.703 0.143 0.545 0.657
=== Confusion Matrix ===
a b <-- classified as
538 32 | a = 0
152 26 | b = 1
4 Heart failure Hasil Akurasi : 76.5886 %
clinical Kappa statistic: 0.4285
records Data MAE: 0.2341
Set RMSE: 0.4839
=== Detailed Accuracy By Class ===
Weighted Avg.
TP Rate FP Rate Precision Recall F-Measure MCC ROC Area PRC Area
0.766 0.363 0.757 0.766 0.757 0.436 0.701 0.693
=== Confusion Matrix ===
a b <-- classified as
179 24 | a = 0
46 50 | b = 1
5 Haberman Hasil Akurasi : 76.143 %
Kappa statistic: 0.2474
MAE: 0.2386
RMSE: 0.4884
=== Detailed Accuracy By Class ===
Weighted Avg.
TP Rate FP Rate Precision Recall F-Measure MCC ROC Area PRC Area
0.761 0.560 0.737 0.761 0.724 0.283 0.601 0.666
=== Confusion Matrix ===
a b <-- classified as
212 13 | a = 1
60 21 | b = 2
3. Jalankan SVM pada semua dataset menggunakan RapidMiner, catatlah hasil eksperimen dalam
sebuah tabel
No Dataset Hasil
1 Iris Hasil Akurasi : 99.33%
Kappa statistic: 0.990
RMSE: 0.026
weighted_mean_precision: 99.44% +/- 1.76% (micro average: 99.35%),
weights: 1, 1, 1
true Iris- true Iris- true Iris- class
setosa versicolor virginica precision
pred. Iris-
50 0 0 100.00%
setosa
pred. Iris-
0 49 0 100.00%
versicolor
pred. Iris-
0 1 50 98.04%
virginica
class recall 100.00% 98.00% 100.00%
2 Wine Hasil Akurasi : 69.08%%
Kappa statistic: 0.524
RMSE: 0.545
weighted_mean_precision: 70.72% +/- 17.30% (micro average: 68.69%),
weights: 1, 1, 1
weighted_mean_recall: 67.44% +/- 14.30% (micro average: 67.41%),
weights: 1, 1, 1
true 1 true 2 true 3
pred. 1 47 6 4
pred. 2 8 53 21
pred. 3 4 12 23
class recall 79.66% 74.65% 47.92%
3 Blood Hasil Akurasi : 76.33%
Transfusion Kappa statistic: 0.142
Service RMSE: 0.416
Center weighted_mean_recall: 55.49% +/- 3.39% (micro average: 55.50%),
weights: 1, 1
weighted_mean_precision: 64.13% +/- 8.25% (micro average: 64.63%),
weights: 1, 1
true 1 true 0 class precision
pred. 1 28 27 50.91%
pred. 0 150 543 78.35%
class recall 15.73% 95.26%
4 Heart failure Hasil Akurasi : 68.23%
clinical Kappa statistic: 0.013
records Data RMSE: 0.469
Set weighted_mean_precision: 39.06% +/- 15.98% (micro average: 84.06%),
weights: 1, 1
weighted_mean_recall: 50.50% +/- 1.58% (micro average: 50.52%),
weights: 1, 1
true 1 true 0 class precision
pred. 1 1 0 100.00%
pred. 0 95 203 68.12%
class recall 1.04% 100.00%
5 haberman Hasil Akurasi : 74.48%
Kappa statistic: 0.187
RMSE: 0.428
weighted_mean_recall: 57.80% +/- 8.09% (micro average: 58.17%),
weights: 1, 1
weighted_mean_precision: 65.19% +/- 16.95% (micro average: 65.70%),
weights: 1, 1
true 1 true 2 class precision
pred. 1 209 62 77.12%
pred. 2 16 19 54.29%
class recall 92.89% 23.46%
