Text Classification with GloVe and GRU

---

Dataset:

• dataset have 5 different class.
• class 1: Love 🧡
• class 2: Sport ⚾
• class 3: Good sentiment 😃
• class 4: Bad sentiment 😔
• class 5: Foods 🍴

Model:

• Use GloVe pre-trained embeddings to convert my text to numerical vectors.
• Use 2 Layer GRU with bidirectional.
• Use 2 Layer fully connected.

Results:

Feature Vector Dimension	Train Loss	Train Accuracy	Validation Loss	Validation Accuracy	Test Loss	Test Accuracy
50d	0.4208	0.85	1.1467	0.64	0.6184	0.80
100d	0.0439	0.99	1.5526	0.64	1.0246	0.79
200d	0.0103	1.	2.1787	0.71	1.3191	0.78
300d	0.0051	1.	1.9680	0.71	1.5144	0.73