# T-SNE数据可视化

[T-分布式随机相邻嵌入(T-SNE)](https://en.wikipedia.org/wiki/T-distributed_stochastic_neighbor_embedding)是由Delft技术大学的Laurens van der Maaten创建的数据可视化工具。

虽然它可以用于任何数据，但是t-SNE（发音为Tee-Snee）只对标记数据有意义，这说明输入是如何聚类的。下面，你可以看到使用t-SNE处理MNIST数据在DL4J中生成的图形类型。

![](https://1666946895-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-LsGrpMiOeoMSFYK0VJQ%2F-LyDIpAWhpYB0StGxCPL%2F-LyDJcNd2GtB8oYI0whD%2Fimage.png?alt=media\&token=2e38bc7b-49a2-45d5-a689-2f25084e4b20)

仔细看，你可以看到数字聚集在它们相似的地方，旁边的点。\
下面是T-SNE如何出现在DL4J中的代码。

```java
public class TSNEStandardExample {

    private static Logger log = LoggerFactory.getLogger(TSNEStandardExample.class);

    public static void main(String[] args) throws Exception  {
        //第1步：初始化
        int iterations = 100;
        //create an n-dimensional array of doubles
        DataTypeUtil.setDTypeForContext(DataBuffer.Type.DOUBLE);
        List<String> cacheList = new ArrayList<>(); //cacheList is a dynamic array of strings used to hold all words

        //第2步：将文本输入转换成单词列表
        log.info("Load & Vectorize data....");
        File wordFile = new ClassPathResource("words.txt").getFile();   //打开文件
        //获取所有唯一词向量的数据
        Pair&lt;InMemoryLookupTable,VocabCache&gt; vectors = WordVectorSerializer.loadTxt(wordFile);
        VocabCache cache = vectors.getSecond();
        INDArray weights = vectors.getFirst().getSyn0();    //将独特词的权重分成自己的列表
        for(int i = 0; i &lt; cache.numWords(); i++)   //把字串分隔成自己的列表
            cacheList.add(cache.wordAtIndex(i));

        //第3步：构建双树TSNE以供以后使用
        log.info("Build model....");
        BarnesHutTsne tsne = new BarnesHutTsne.Builder()
                .setMaxIter(iterations).theta(0.5)
                .normalize(false)
                .learningRate(500)
                .useAdaGrad(false)
//                .usePca(false)
                .build();

        //第4步：建立TSNE值并将其保存到文件中
        log.info("Store TSNE Coordinates for Plotting....");
        String outputFile = "target/archive-tmp/tsne-standard-coords.csv";
        (new File(outputFile)).getParentFile().mkdirs();
        tsne.plot(weights,2,cacheList,outputFile);
      //这个tsne将使用向量的权重作为它的矩阵，具有两个维度，使用单词字符串作为标签，并将其写入在前一行创建的outputFile

    }



}
```

![](https://firebasestorage.googleapis.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-LsGrpMiOeoMSFYK0VJQ-1479650856%2Fuploads%2FYUQVIYtzDG8Ymp2iVZwX%2Ffile.gif?alt=media)​这是使用gnuplot绘制的tsne-standard-coords.csv文件的图像。

![Tsne data plot](https://1666946895-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2Fdeeplearning4j%2F-LulFhOUUvvDel8FTBAl%2F-LulG94vX8cEDEVww4ye%2Ftsne_output.png?generation=1574932126186134\&alt=media)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://deeplearning4j.konduit.ai/zhong-wen-v1.0.0/tiao-you-yu-xun-lian/tsne-shu-ju-ke-shi-hua.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.