リフォーム年月日
地上の居住エリアの面積（平方フィート）
地下室の合計面積（平方フィート）
ガレージの面積（平方フィート）
The first component, PC1, seems to be a kind
of "size" component,
all of the features have the same sign (positive),
indicating that this component is describing
a contrast between houses having large values
and houses having small values for these features.
Principal Component Analysis. 相関の強い観測変数を統合して、新たな合成変数を作成する分類のための分析手法です。 多数の量的変数を少数の量的変数に縮約する目的で利用されます。
The interpretation of the third component PC3 is a little trickier. The features GarageArea and YearRemodAdd both have near-zero loadings, so let's ignore those. This component is mostly about TotalBsmtSF and GrLivArea. It describes a contrast between houses with a lot of living area but small (or non-existant) basements, and the opposite: small houses with large basements.
The interpretation of the third component PC3 is a little trickier.
The features GarageArea and YearRemodAdd both have near-zero
loadings, so let's ignore those.
This component is mostly about TotalBsmtSF and GrLivArea.
It describes a contrast between houses with a lot of living area but small basements,
and the opposite: small houses with large basements.
The interpretation of the third component PC3 is a little trickier. The features GarageArea and YearRemodAdd both have near-zero loadings, so let's ignore those. This component is mostly about TotalBsmtSF and GrLivArea. It describes a contrast between houses with a lot of living area but small (or non-existant) basements, and the opposite: small houses with large basements.
Your goal in this question is to use the results of PCA to
discover one or more new features that improve the performance of your model.
One option is to create features inspired by the loadings, like we did in the tutorial.
Another option is to use the components themselves as features
(that is, add one or more columns of X_pca to X).
The next question explores a way you can use PCA
to detect outliers in the dataset
(meaning, data points that are unusually extreme in some way).
Outliers can have a detrimental有害 effect on model performance,
so it's good to be aware of them in case you need to take corrective action.
PCA in particular can show you anomalous異常な variation which might not be apparent from the original features:
neither small houses nor houses with large basements are unusual,
but it is unusual for small houses to have large basements.
That's the kind of thing a principal component can show you.
As you can see,
in each of the components
there are several points
lying at the extreme ends of
the distributions -- outliers, that is.
3) Outlier Detection
Do you notice any patterns in the extreme values?
Does it seem like the outliers are coming from some special subset of the data?
Notice that there are several dwellings listed as Partial sales 部分売却
in the Edwards neighborhood that stand out.
A partial sale is what occurs when there are multiple owners of a property
and one or more of them sell their "partial" ownership of the property.
These kinds of sales are often happen
during the settlement of a family estate or the dissolution of a business
家族の遺産整理や事業の解散の際
and aren't advertised publicly.
If you were trying to predict the value of a house on the open market,
you would probably be justified in removing sales like these from your dataset
-- they are truly outliers.