Four points on a number line (0, 1, 4, 6) are merged step by step into clusters, comparing single, complete, and average linkage rules. Three dendrograms show how each rule produces a different final merge height (3, 6, and 4.5) from the same data. The animation clarifies that dendrogram height encodes distance while leaf position does not, then applies the idea to a learning-analytics example, stressing that clusters reflect data patterns, not personal judgments about students. Useful for first-year AI students learning unsupervised learning fundamentals.
Narrated · 16:9 · Preview before teaching · automatic layout checks do not establish subject accuracy
# Different linkage rules produce different clustering trees Place four points at 0, 1, 4, and 6 on a number line. Begin agglomerative clustering with one cluster per point. Merge 0 and 1 at distance 1, then merge 4 and 6 at distance 2. Compare three definitions for the final intercluster distance: single linkage gives 3, complete linkage gives 6, and average linkage gives (4+6+3+5)/4 = 4.5. Build three dendrograms whose final heights match these values. Explain that dendrogram height represents linkage distance, while horizontal leaf placement is not a numerical coordinate. Cut each tree horizontally to obtain clusters. Finish with a brief learning-support example using synthetic activity patterns, emphasizing that clusters describe data relationships rather than student ability, character, or personal worth. Use a black background, clean tree branches, and consistent colored clusters. Useful for understanding both hierarchical clustering and responsible interpretation. Audience: 一年级人工智能专业学生