R 出错：使用 hclust() 调用国外 function 中的 NA/NaN/Inf

Question

我试图通过开始投手的名字来聚集 MLB 数据。 我梳理了我正在使用的数据，没有任何值为 na 的数据，我在下面的代码中省略了它们。 Clustdata 对我来说看起来非常好ClustData 预览，但我收到此错误：

由 hclust(d, method = "single", members = clustdata[, 1]) 中的 coercionError 引入的 NA：外国 function 调用中的 NA/NaN/Inf (arg 7)

我想通过这些属性按投手名称对该表进行聚类有人有什么想法吗？ 谢谢！ 我是 R 的新手

data7 = read.csv("GL2007.csv", header = T)

data8 = data.frame(na.omit(data7[c(10,23,24,25,26,30,31,33,105)]))
scoreagg = aggregate(v_score ~ h_starting_pitcher_name, data8, mean)
hitsagg = aggregate(v_hits ~ h_starting_pitcher_name, data8, mean)
doubagg = aggregate(v_doubles~ h_starting_pitcher_name, data8, mean)
tripagg = aggregate(v_triples~ h_starting_pitcher_name, data8, mean)
hragg = aggregate(v_homeruns ~ h_starting_pitcher_name, data8, mean)
hbpagg = aggregate(v_hit_by_pitch ~ h_starting_pitcher_name, data8, mean)
walksagg = aggregate(v_walks~ h_starting_pitcher_name, data8, mean)
SOagg = aggregate(v_strikeouts~ h_starting_pitcher_name, data8, mean)

clustdata = data.frame(scoreagg$h_starting_pitcher_name, scoreagg$v_score,hitsagg$v_hits,doubagg$v_doubles,tripagg$v_triples,hragg$v_homeruns,hbpagg$v_hit_by_pitch,walksagg$v_walks,SOagg$v_strikeouts)


library(NbClust)
d = dist(as.matrix(clustdata[,2:9]), method = "euclidean")
hc_1 = hclust(d, method = "single", members = clustdata[,1])

Answer 1

由于问题中没有提供很多细节，因此您似乎没有正确使用members参数。

如果您的目标只是获得聚类，请将其保留为 NULL。

hc_1 = hclust(d, method = "single")