what does the dataset look like?
#Insurance regression model help
analog panther
analog panther
if they are all categorical try this code and see if it works at all. It will likely be worse than XGBoost and Random forest but its worth a try with the little info i have so far:
import tensorflow as tf
class try_model(tf.keras.Model):
def init(self):
self.norm = tf.keras.layers.BatchNormalization()
self.intermediate_layers = [tf.keras.layers.Dense(128, activation="relu") for _ in range(3)]
self.out_layer = tf.keras.layers.Dense(1)
def call(self, x):
x = self.norm(x)
for i in self.intermediate_layers:
x = i(x)
return self.out_layer(x)
model = try_model()
model.fit(x_train, y_train, validation_split=0.2, epochs=100, callbacks=[tf.keras.callbacks.EarlyStopping(patience=15, restore_best_weights=True)])
worthy nebula
how bad is it? maybe its the preprocessing?