Skip to content

Latest commit

 

History

History
42 lines (36 loc) · 1.97 KB

README.md

File metadata and controls

42 lines (36 loc) · 1.97 KB

Reuters Text Classification

We use the Reuters-21578 Text Categorization Collection to perform multi-label text classification with Transformer-based models. Given a Reuters news article, we need to classify it into one or more topic.

The dataset is imbalanced, and contains data of variable length (as illustrated by the tables below). As such, we leverage methods like Inverse Class Frequency to address the dataset imbalance.

Distribution of Tokens (using bert-base-cased):

Metric Value
Average 121.95
Standard deviation 109.25
Maximum 2976
Minimum 4

Distribution of Topics:

Topic Count
[earn] 3687
[acq] 1994
[crude] 326
[trade] 307
[money-fx] 243
... ...

Results

BERT

After finetuning the last two layers of a BERT-cased model + an added dense layer for 22 epochs (refer to train_bert.py), we have the following Validation Data result:

Metric Value
Exact-Match Accuracy 0.9065
Hamming Loss 0.0026
Micro-F1 0.9426
Macro-F1 0.833

XLNet

After finetuning the last two layers of a XLNet-cased model + an added dense layer for 12 epochs (refer to train_xlnet.py), we have the following Validation Data result:

Metric Value
Exact-Match Accuracy 0.923
Hamming Loss 0.0019
Micro-F1 0.956
Macro-F1 0.905