Docs: Does not reflect the current release of ML.Net

[PeterPann23]

Code sample does not reflect the API of the current NuGet package.

---

Document Details

⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.

• ID: 925c8fe9-ba0b-bd2c-cea0-9d851ece8825
• Version Independent ID: 41cafff5-6418-2e54-3a0d-ee29fb1fca1f
• Content: Cluster iris flowers using a clustering learner - ML.NET
• Content Source: docs/machine-learning/tutorials/iris-clustering.md
• Product: dotnet-ml
• GitHub Login: @pkulikov
• Microsoft Alias: johalex

[pkulikov]

@PeterPann23 what code sample do you refer to?

The sample project uses ML.NET v0.10; though it might skip using some latest additions.

[PeterPann23]

Hi pkulikov
I think that the Iris sample should be using the new API when mapping the class to the columns.

[pkulikov]

@PeterPann23 do you mean the IrisData class should use the LoadColumn attribute? That is, apply the approach from the issue classification tutorial. If yes, I think it's a good suggestion and I'll update the tutorial the next week.

@JRAlexander the price prediction tutorial also might use the LoadColumn attribute as the current usage of the Column attribute only specifies the source index and doesn't define the name of a column (which is not the case for the sentiment analysis tutorial).

[PeterPann23]

Yes, that's what I was thinking about, also to use the new method ( MLContext.Data.ReadFromTextFile) that will actually throw you a null exception if you not add the annotation. Some annotations are needed for one and not for the other (textloaderarguments). A cleanup of the concept may be needed but that's a different topic.

I feel for those that get introduced one should not learn the "old" way as well as understand the implications….

[pkulikov]

@PeterPann23 thanks for raising the issue; I've created two PRs that simplify data loading step as a result.

[pkulikov]

@PeterPann23 also it looks that the old API is going to be removed: dotnet/machinelearning#2593