---
language:
  - en
multilinguality:
  - monolingual
license: other
license_name: topicnet
license_link: >-
  https://github.com/machine-intelligence-laboratory/TopicNet/blob/master/LICENSE.txt
task_categories:
  - text-classification
task_ids:
  - topic-classification
  - multi-class-classification
  - multi-label-classification
tags:
  - topic-modeling
  - topic-modelling
  - text-clustering
  - multimodal-data
  - multimodal-learning
  - modalities
  - document-representation
---


# Stanford Triplets

Some measurable characteristics of the dataset:

* D — number of documents
* <modality name> W — modality dictionary size (number of unique tokens)
* <modality name> len D — average document length in modality tokens (number of tokens)
* <modality name> len D uniq — average document length in unique modality tokens (number of unique tokens)


|       |    D |   @word W |   @word len D |   @word len D uniq |
|:------|------------:|-----------------:|---------------------:|--------------------------:|
| value |      963564 |      3.55391e+09 |               3688.3 |                   962.404 |


Information about document lengths in modality tokens:


|      |   len_total@word |   len_uniq@word |
|:-----|-----------------:|----------------:|
| mean |           3688.3 |         962.404 |
| std  |           3482.8 |         458.434 |
| min  |              0   |           0     |
| 25%  |           1791   |         665     |
| 50%  |           2835   |         878     |
| 75%  |           4608   |        1184     |
| max  |         754814   |       17874     |


**Metadata**: There are triplets that can be used to measure quality.