阅读量:1
在SciPy中,可以使用scipy.spatial.distance
模块来计算两个样本集之间的距离和相似性。
- 计算距离:
from scipy.spatial import distance # 样本集合1 X = [[0, 1], [1, 1]] # 样本集合2 Y = [[2, 0], [3, 1]] # 计算欧氏距离 euclidean_distance = distance.cdist(X, Y, 'euclidean') print("Euclidean Distance:") print(euclidean_distance) # 计算曼哈顿距离 manhattan_distance = distance.cdist(X, Y, 'cityblock') print("\nManhattan Distance:") print(manhattan_distance) # 计算闵可夫斯基距离 minkowski_distance = distance.cdist(X, Y, 'minkowski', p=2) print("\nMinkowski Distance:") print(minkowski_distance)
- 计算相似性:
from scipy.spatial import distance # 样本集合1 X = [[0, 1], [1, 1]] # 样本集合2 Y = [[2, 0], [3, 1]] # 计算余弦相似度 cosine_similarity = 1 - distance.cdist(X, Y, 'cosine') print("Cosine Similarity:") print(cosine_similarity) # 计算皮尔逊相关系数 pearson_similarity = 1 - distance.cdist(X, Y, 'correlation') print("\nPearson Correlation:") print(pearson_similarity)
以上代码分别展示了如何计算欧氏距禮、曼哈顿距离、闵可夫斯基距离、余弦相似度和皮尔逊相关系数。可以根据具体需求选择合适的距离或相似性度量方法。