特徴抽出を使ってテキスト検索ツールを作成する
「不思議の国のアリス」の原作(英語)からの文章で特徴検出器を訓練する.
In[1]:=
![Click for copyable input](assets.ja/create-a-text-search-tool-using-feature-extraction/In_16.png)
alice = ExampleData[{"Text", "AliceInWonderland"}];
sentences = TextSentences[alice];
fe = FeatureExtraction[sentences]
Out[1]=
![](assets.ja/create-a-text-search-tool-using-feature-extraction/O_11.png)
データ集合の抽出された特徴についてNearestFunctionを生成する.
In[2]:=
![Click for copyable input](assets.ja/create-a-text-search-tool-using-feature-extraction/In_17.png)
nf = Nearest[fe[sentences] -> Automatic]
Out[2]=
![](assets.ja/create-a-text-search-tool-using-feature-extraction/O_12.png)
「不思議の国のアリス」中の最も近い文を表示する関数を構築し,これをいくつかの文に使ってみる.
In[3]:=
![Click for copyable input](assets.ja/create-a-text-search-tool-using-feature-extraction/In_18.png)
nearestalice = sentences[[First@nf[fe[#]]]] &;
In[4]:=
![Click for copyable input](assets.ja/create-a-text-search-tool-using-feature-extraction/In_19.png)
nearestalice["Alice and the Rabbit"]
Out[4]=
![](assets.ja/create-a-text-search-tool-using-feature-extraction/O_13.png)
In[5]:=
![Click for copyable input](assets.ja/create-a-text-search-tool-using-feature-extraction/In_20.png)
nearestalice["Off her head"]
Out[5]=
![](assets.ja/create-a-text-search-tool-using-feature-extraction/O_14.png)