LSTM based Time-Series Prediction Model in R

I need a trained, validated and tested model with LSTM to predit the output.
One year long data with 4 to 5 (max) input parameters and 1 output ..data is with 15 minutes frequency. The model is intended to be used for development of preditice control. Please contact only, if you have advanced expertise in R..

Some important information>
1- Pre-preocessing of data to check null values, correlation and seasonal variations in the input/output
2- Training with 80% data, validation with 15% data and testing with 5% data (with 15 minutes time step)
3- Model performnace evaluation indeicators like MAPE, MAD, MSE (evolution of perfromnace indicators with time)
4- If Normalised, data should be converted to orignal format for plotting the results of precite v/s actual and plotted against the date & time
5- Sensitivity Analysis of the Input factors using "NeuralSens" package of R and with dervative based method..
6- Model Preditive Control (MPC) control algorithim....
7- The structure of the LSTM woul be required alongwith a short descriptive report...

You can contact for any