CSCI_5080_Assignment_07_1

CSCI_5080_Assignment_07 Question 1 A 1(2, 10), A 2(2, 5), A 3(8, 4), B 1(5, 8), B 2(7, 5), B 3(6, 4), C 1(1, 2), C 2(4, 9). Using R to solve for the Euclidean distance > data_hmwk7 <- matrix(c(2,10,2,5,8,4,5,8,7,5,6,4,1,2,4,9), byrow = T, nrow = 8 ) > hmwk7<- dist(data_hmwk7, method = "euclidean") > hmwk7 1 2 3 4 5 6 7 2 5.000000 3 8.485281 6.082763 4 3.605551 4.242641 5.000000 5 7.071068 5.000000 1.414214 3.605551 6 7.211103 4.123106 2.000000 4.123106 1.414214 7 8.062258 3.162278 7.280110 7.211103 6.708204 5.385165 8 2.236068 4.472136 6.403124 1.414214 5.000000 5.385165 7.615773 Dataset A1(2, 10) B1(5, 8) C1(1, 2) Cluster Assignment A1 (2,10) 0 3.605551 8.06 1 A2 (2,5) 5 4.24 3.16 3 A3 (8,4) 8.485281 5 7.27 2 B1 (5,8) 3.605551 0 7.211103 2 B2 (7,5) 7.071068 3.605551 6.708204 2 B3 (6,4) 7.211103 4.123106 5.385165 2 C1 (1,2) 8.062258 7.211103 0 3 C2 (4,9) 2.236068 1.414214 7.615773 2 (a.) After the initial round of execution, the clustering result comprises three clusters: (1) consisting of {A1}, (2) including {B1, A3, B2, B3, C2}, and (3) containing {C1, A2}. The centers for these clusters are derived by averaging the x- values and y-values, resulting in coordinates (1) (2, 10), (2) (6, 6), and (3) (1.5, 3.5). (b.) 2nd session 1 CSCI_5080_Assignment_07

Dataset Cluster1(2, 10) Cluster2(6, 6) Cluster 3(1.5, 3.5) Cluster Assignment A1 (2,10) 0 5.66 6.52 1 A2 (2,5) 5 4.123 1.58 3 A3 (8,4) 8.485281 2.83 6.52 2 B1 (5,8) 3.605551 2.23 5.7 2 B2 (7,5) 7.071068 1.414 5.7 2 B3 (6,4) 7.211103 2 4.53 2 C1 (1,2) 8.062258 6.4 1.58 3 C2 (4,9) 2.236068 3.6056 6.04 1 (1) {A1, C2} (2) {A3, B1, B2, B3}, (3) {A2, C1} with centers (1) (3, 9.5), (2) (6.5, 5.25), (3) (1.5, 3.5) The Third Session Datase t Cluster_c1( 3, 9.5) Cluster_c2(6. 5, 5.25) Cluster_c 3(1.5, 3.5) Assignme nt of Cluster A1 (2,10) 1.118 6.543 6.52 1 A2 (2,5) 4.61 4.51 1.58 3 A3 (8,4) 7.433 1.95 6.52 2 B1 (5,8) 2.5 3.13 5.7 1 B2 (7,5) 6.02 0.56 5.7 2 B3 (6,4) 6.26 1.35 4.53 2 C1 (1,2) 7.76 6.388 1.58 3 C2 (4,9) 1.11 4.5 6.04 1 The final three clusters are (1) {A1, C2, B1}, (2) {A3, B2, B3}, (3) {C1, A2} Question 2: The Clustering-Based SVM (CB-SVM) is specifically crafted to address challenges posed by large datasets, where traditional Support Vector Machines (SVMs) may underperform when trained on the entire dataset. To mitigate this, CB-SVM employs a hierarchical micro-clustering algorithm, scanning the entire dataset only once to provide the SVM with high-quality samples containing statistical summaries. This approach maximizes the 2 CSCI_5080_Assignment_07

3. Decluster entries near the boundary into the next level, accumulating children entries declustered from parent entries into the training set with non-declustered parent entries. 4. Construct another SVM from centroids of entries in the training set and repeat from step 3 until no further accumulation occurs. Time spent: - 2hrs. 4 CSCI_5080_Assignment_07

PART 2 Lesson 3: 5 CSCI_5080_Assignment_07

Dependency Network Tab: 7 CSCI_5080_Assignment_07

TM_Clustering Model: Cluster Diagram Tab Cluster Profiles Tab Cluster Characteristics Tab 8 CSCI_5080_Assignment_07

Attribute Profile Tab: Attribute Characteristics Tab: 10 CSCI_5080_Assignment_07

Attribute Discrimination Tab: Time spent: 4 hours. 11 CSCI_5080_Assignment_07

> # 3rd cluster > dist(rbind(b,a1), method="euclidean") b a1 6.519202 > dist(rbind(b,a2), method="euclidean") b a2 1.581139 > dist(rbind(b,a3), method="euclidean") b a3 6.519202 > dist(rbind(b,a4), method="euclidean") b a4 5.700877 > dist(rbind(b,a5), method="euclidean") b a5 5.700877 > dist(rbind(b,a6), method="euclidean") b a6 4.527693 > dist(rbind(b,a7), method="euclidean") b a7 1.581139 > dist(rbind(b,a8), method="euclidean") b a8 6.041523 > # 3rd Section > c1=c(3, 9.5) > c2=c(6.5, 5.25) > c3=c(1.5, 3.5) > # > dist(rbind(c1,a1), method="euclidean") c1 a1 1.118034 > dist(rbind(c1,a2), method="euclidean") c1 a2 4.609772 > dist(rbind(c1,a3), method="euclidean") c1 a3 7.433034 > dist(rbind(c1,a4), method="euclidean") c1 a4 2.5 > dist(rbind(c1,a5), method="euclidean") c1 a5 6.020797 > dist(rbind(c1,a6), method="euclidean") c1 a6 6.264982 > dist(rbind(c1,a7), method="euclidean") c1 a7 7.762087 > dist(rbind(c1,a8), method="euclidean") 13 CSCI_5080_Assignment_07

c1 a8 1.118034 > # > dist(rbind(c2,a1), method="euclidean") c2 a1 6.543126 > dist(rbind(c2,a2), method="euclidean") c2 a2 4.506939 > dist(rbind(c2,a3), method="euclidean") c2 a3 1.952562 > dist(rbind(c2,a4), method="euclidean") c2 a4 3.132491 > dist(rbind(c2,a5), method="euclidean") c2 a5 0.559017 > dist(rbind(c2,a6), method="euclidean") c2 a6 1.346291 > dist(rbind(c2,a7), method="euclidean") c2 a7 6.388466 > dist(rbind(c2,a8), method="euclidean") c2 a8 4.506939 > # > dist(rbind(c3,a1), method="euclidean") c3 a1 6.519202 > dist(rbind(c3,a2), method="euclidean") c3 a2 1.581139 > dist(rbind(c3,a3), method="euclidean") c3 a3 6.519202 > dist(rbind(c3,a4), method="euclidean") c3 a4 5.700877 > dist(rbind(c3,a5), method="euclidean") c3 a5 5.700877 > dist(rbind(c3,a6), method="euclidean") c3 a6 4.527693 > dist(rbind(c3,a7), method="euclidean") c3 a7 1.581139 > dist(rbind(c3,a8), method="euclidean") c3 a8 6.041523 14 CSCI_5080_Assignment_07