From: Celebrity profiling through linguistic analysis of digital social networks
Parameter | Value |
---|---|
Model | Keras sequential - Tensorflow |
Loss | sparse_categorical_crossentropy |
Optimizer | Adam |
Epochs | 5 |
Batch size | 10 |
Activation input layer | ReLu |
Activation output layer | Softmax |
Dense input layer | 10 |
Dense output layer | 3, 8, 70 |
Features (proposed features) | 1365 |
Features (lexical features) | 1000 |