how to get the indices of specific set of rows

Question

feature_log_prob_sorted = np.abs(naive.feature_log_prob_).argsort()

feature_log_prob_sorted = feature_log_prob_sorted[ : : -1]

Here naive is BernoulliNB's object and feature_log_prob_ is empirical log probability of features given a class.

After applying the NB model on certain data I was trying to get the top 20 probabilities.

argsort() function return the indices of values from low probability to high probability so I have inverted the data.

Now I want to know the indices of top 20 value. How can I get it?

Answer 1

要获取排序列表的前n值，请使用：

top_n = feature_log_prob_sorted[:n]