Q34. A student writes:

kmeans = KMeans(n_clusters=3)
df["cluster"] = kmeans.fit(X)

What is the issue?

Question

Q34. A student writes:

kmeans = KMeans(n_clusters=3)
df["cluster"] = kmeans.fit(X)

What is the issue?

BlackTom AI · Accepted Answer

fit() in scikit-learn returns the estimator object (self) after fitting; it does not return an array of cluster labels. Use fit_predict(X) or access the labels_ attribute to get the labels.

1) Relevant concepts
- scikit-learn estimator API: .fit(X[, y]) fits the model and returns the estimator (self).  
- For clustering, KMeans provides .fit_predict(X) which fits and returns the cluster labels (an array of length n_samples). After fitting, the labels are available as the attribute estimator.labels_. .predict(X_new) assigns clusters for new data using the fitted model.

2) Step-by-step reasoning
- The student wrote: kmeans = KMeans(n_clusters=3) followed by df["cluster"] = kmeans.fit(X)
- Step A: kmeans.fit(X) trains the KMeans model on X and returns the fitted KMeans object (not labels).
- Step B: Assigning that returned object to df["cluster"] attempts to put the estimator into a DataFrame column, which is wrong (it’s not the desired labels array and will either store the object or raise an error on shape mismatch).
- Correct approaches:
  - labels = kmeans.fit_predict(X); df["cluster"] = labels
  - kmeans.fit(X); df["cluster"] = kmeans.labels_

3) Intermediate “execution” steps (conceptual)
- Call kmeans.fit(X) -> returns KMeans instance (contains centroids_, labels_, inertia_)
- To get labels after fit: labels = kmeans.labels_ (array shape = (n_samples,))
- Or combine: labels = kmeans.fit_predict(X) -> labels directly

4) Why the correct answer is correct
Because .fit() follows scikit-learn’s API and returns the estimator, not a labels array. fit_predict and labels_ are the designed ways to access cluster assignments.

5) Why other options are incorrect
- "KMeans cannot be used with DataFrames": incorrect — KMeans accepts array-like inputs including NumPy arrays and pandas DataFrames (converted internally).
- "The number of clusters must be 2": incorrect — you can choose any sensible integer >=1; here n_clusters=3 is valid.
- "The DataFrame must be sorted first": irrelevant to KMeans fitting.
- "The model must be saved before fitting": nonsensical — you fit (train) before saving; saving is optional.

6) Conclusion
The issue is that .fit() does not return cluster labels; use .fit_predict() or the .labels_ attribute. This matches the provided correct answer.

Q34. A student writes: kmeans = KMeans(n_clusters=3) df["cluster"] = kmeans.fit(X) What is the issue?单项选择题

类似问题

The goal of clustering a set of data is to:

Which of the following statements about clustering in machine learning is true?

Your team experiments with increasing K in K-Means from 3 to 8. What is the expected effect?

In k-means clustering, what is the primary goal of the algorithm? (you can choose more than one)

The elbow plot helps determine the optimal number of clusters by looking for:

What does clustering accomplish?

Which method is best for detecting patterns in unlabeled data?

Which of the descriptions listed below is TRUE?

更多留学生实用工具

智能学习助手

风格化写作助手

论文查重助手

文献引用助手

课堂转译助手

课堂笔记助手

Quiz搜索助手

学校历年真题

智能刷题助手

智能匹配练习

希望你的学习变得更简单