Table 4 Performance of the mutual information feature selection on the StockTwits dataset
From: Big Data: Deep Learning for financial sentiment analysis
Features | Accuracy | Precision | Recall | F-measure | AUC |
|---|---|---|---|---|---|
55,820 | 0.7088 | 0.7134 | 0.6980 | 0.7056 | 0.7088 |
40,000 | 0.5417 | 0.5311 | 0.7115 | 0.6082 | 0.5417 |
20,000 | 0.5123 | 0.5087 | 0.7144 | 0.5943 | 0.5123 |
4000 | 0.5391 | 0.5337 | 0.6190 | 0.5732 | 0.5391 |
2000 | 0.5406 | 0.5350 | 0.6193 | 0.5741 | 0.5406 |
400 | 0.5665 | 0.5540 | 0.6815 | 0.6112 | 0.5665 |
200 | 0.4713 | 0.4760 | 0.5692 | 0.6126 | 0.5459 |
50 | 0.5077 | 0.5052 | 0.7414 | 0.6009 | 0.5077 |