jfpuget / libfm_in_keras Goto Github PK

hi，
as we can see from the codes:
i):features = ['os', 'device', 'app']
f_size = [int(data[f].max()) + 1 for f in features]

ii) def get_embed(x_input, x_size, k_latent):
if x_size > 0: #category
embed = Embedding(x_size, k_latent, input_length=1,
embeddings_regularizer=l2(embedding_reg))(x_input)
embed = Flatten()(embed)
else:
embed = Dense(k_latent, kernel_regularizer=l2(embedding_reg))(x_input)
return embed

iii) biases = [get_embed(x, size, 1) for (x, size) in zip(input_x, f_size)]
factors = [get_embed(x, size, k_latent) for (x, size) in zip(input_x, f_size)]

in the demo,there are all the category variable in the data,
can you show us how to handle with continuous variable?or make continuous variable into bins?
looking for your reply!
^_^

Hi,
Many thanks sharing this code.
Having the output file, how is it possible to connect it to the original dataset ? I'm using panda, is there any special command doing it ? Appear to me that the usual "merge on" command is not working with multiple columns