I am working on developing a predictive drive cycle for a specific city. I have gathered driving characteristics parameters for my target city. I have obtained similar data from previous research works and their corresponding drive cycles. I would like to train a network using the driving characteristic data obtained from the research papers keeping the time stamped drive cycle obtained from the corresponding researches as target so that the network can predict a drive cycle based on my collected data. I have experience of working on ML using TensorFlow and Keras.
I think it would be appropriate to try a solution with RNN. Note that the standard RNN has a problem with the so-called fading gradient. I recommend using the LSTM network on a non-sequential input data set, it is enhanced by introducing an additional closed and cellular state.
Have a look at https://towardsdatascience.com/3-types-of-sequence-prediction-problems-97f22e946318
https://www.datacamp.com/community/tutorials/deep-learning-python
I think it would be appropriate to try a solution with RNN. Note that the standard RNN has a problem with the so-called fading gradient. I recommend using the LSTM network on a non-sequential input data set, it is enhanced by introducing an additional closed and cellular state.
Prof. Nikolay Raychev thanks for your input sir. I have already done a fair amount of experimentation using simple RNNs, LSTM and GRU as well. But in all the cases I have tried out so far, both input and output data were sequential, for example, a character level language modelling. But in the problem I am trying to figure out, only the output data is sequential. The inputs are numeric in nature. Selections of layers are confusing me. Do you think using a dense layer at input side and using a RNN layer at the output will do the trick?
The inputs are numeric in nature.
Selections of layers are confusing me. Do you think using a dense layer at input side and using a RNN layer at the output
Maha Abd Alellah I'm trying to solve it with that type of a configuration. I have experience of NLP works using RNN. For example, the IMDb dataset, the output is numeric and input is sequential. I have implemented LSTM layer at the input side and Dense layer at the output. The problem I have in hand is kind of a reverse architecture of the IMDb sentiment analysis in a way I guess.
- Badges
- Science topic