The entire project is fixed rate. I have certain milestones that you should accomplish. Entire project is contingency.
We will proceed to other milestones only when You complete the Milestone 1 with a novel approach and with better performance than other models (or at least be in the top 4 RMSE scores for milestone 1).
Phase 1: Devise new ML/Deep learning strategy for the use case with preferable implementation in python keras. The strategy applied here should have the capability to result in significant contribution to this use case. The model devised in this phase should not been applied previously in related area and should be equivalent to a research work.
Aim to get the score equal to the top scores and the score of your model should atleast score top 2 or 3 in RMSE score metric and score function provided by author. Graph of predicted RUL and actual RUL should almost same proving that the model is best one. Provide a brief report on why the model is best compared to other models. Why this model has been chosen. Provide the comparison with other models tried.
Test the accuracy on 4 other similar data sets and graphs of actual vs predicted.
August 4, 2018
I am looking for a mix of experience and value