I hope you’re all doing well!
I’m currently working on a exciting open-source project that involve Node.js and TensorFlow.js, and I’m looking to extent the features on my code.
Current Implementation:
The API I’ve built allows for:
- Collecting Measurement Data: The API can collect and process measurement data.
- Building Custom ML Models: It can build models with adjustable hyper-parameters.
- Training Data: The API can train the collected data using the custom models.
- Classifying Measurements: It can classify unknown measurements based on the trained model. For example, if you request to classify -5100.00, the API returns the class “Negative”. If you request 23.10, the API returns the class “Positive” - useless isn’t it but as a PoC it work well.
What I Need Help With:
While the current implementation works well as a proof of concept, I need help to extend the capabilities of the API to predict time series data instead of just classification. Specifically:
- Enhancing the Model: Adjusting the current model to handle time series predictions.
- Implementing LSTM and Other Networks: Setting up the model to use LSTM or other appropriate networks for time series.
- Node.js and TensorFlow.js Expertise: I need guidance and support on the implementation as I’m a beginner and not entirely confident in my current code, even though it has produced good results so far.
The reason I’m asking such support is because I’m not a technical expert and I believe I’m very close to what I expect as end result but couldn’t achieve my goal (alone).
If you’re interested on spending time to teach me, review my code or pair-program, please drop a comment below or send me a direct message. Looking forward to collaborating and building something amazing together!
Thank you!