Time Series Forecasting with LSTM & AI
Budget: $250 – $750 USD
(PLEASE READ THE PROJECT ASKS QUESTIONS WHICH ARE NOT MENTIONED HERE, DO NOT START A THREAD TO REGNOTIATE THE PRICE WHICH YOU HAVE QUOTED, SO PLEASE QUOTE CORRECTLY)
MileStones:
* 10% at the time of awarding.
* 25% at the time of showing a demo (after understanding our output)
* 15% When the completion is done at your infra.
* 25% You install the same.. and make it run.
* 25% Once all things are working and you hand over.
BONUS
* 15% if you say a time and deliver on time.
I am seeking a data scientist proficient in time series analysis and machine learning techniques. The objective of this project is to predict future values based on our internal generated data using various AI models.
More about the project:
I have an API, which allows me pull json / XML record (each record has over 400 Variables).
Those records would also have two "result" values X,Y (so each record has an X and Y, which is derived from above variables)
We would have 1000 and more of these record, all are sequential.
We would want you to install multiple models to predict X and Y.
We would give you Server for you to host for us, you would have to use Postgres DB.
Requirements
(a) We would like to add records at our own will, and the model should be able to train using new and older data to get the results.
Exampe, we start with say 1000 records, and we want you predict next 5 X,Y values, then we add 100 more record, we would like to get next 5 X,Y values with the 1100 rows of data.
(B) we would like to retrain the model.
(C) As there are different models, we would like to see results in different models, so that we could see which model is predicting based on the values we want.
(d) we could give you all the data in API, so that you could use the same in the model
(e) You would have to tell us how much time it would take to retrain.
(f) We would give you an API, so that you could generate and push the prediction data in db.
(g) We have multiple sets (each set of 1000 records), and we would like, so each set has to be trained differently to predict the right data.
(h) We would like to add more variables as we go along (400 to 500), we would then retrain the model to 500 variables to get the data.
(i) We would need the Source code, and instruction to recompile the software as well and installation guide.
Key Tasks:
- Utilize LSTM to forecast time series data.
- Implement ARIMA, GRU, and Prophet for comparative analysis and validation.
- Generate accurate predictions from our internal data.
Ideal Skills:
- Expertise in LSTM and other AI models.
- Proficiency in time series data analysis.
- Experience with ARIMA, GRU, and Prophet.
More about the project
I look forward to your bids. Please include your relevant experience and previous projects you have completed in this field.
MileStones:
* 10% at the time of awarding.
* 25% at the time of showing a demo (after understanding our output)
* 15% When the completion is done at your infra.
* 25% You install the same.. and make it run.
* 25% Once all things are working and you hand over.
BONUS
* 15% if you say a time and deliver on time.
I am seeking a data scientist proficient in time series analysis and machine learning techniques. The objective of this project is to predict future values based on our internal generated data using various AI models.
More about the project:
I have an API, which allows me pull json / XML record (each record has over 400 Variables).
Those records would also have two "result" values X,Y (so each record has an X and Y, which is derived from above variables)
We would have 1000 and more of these record, all are sequential.
We would want you to install multiple models to predict X and Y.
We would give you Server for you to host for us, you would have to use Postgres DB.
Requirements
(a) We would like to add records at our own will, and the model should be able to train using new and older data to get the results.
Exampe, we start with say 1000 records, and we want you predict next 5 X,Y values, then we add 100 more record, we would like to get next 5 X,Y values with the 1100 rows of data.
(B) we would like to retrain the model.
(C) As there are different models, we would like to see results in different models, so that we could see which model is predicting based on the values we want.
(d) we could give you all the data in API, so that you could use the same in the model
(e) You would have to tell us how much time it would take to retrain.
(f) We would give you an API, so that you could generate and push the prediction data in db.
(g) We have multiple sets (each set of 1000 records), and we would like, so each set has to be trained differently to predict the right data.
(h) We would like to add more variables as we go along (400 to 500), we would then retrain the model to 500 variables to get the data.
(i) We would need the Source code, and instruction to recompile the software as well and installation guide.
Key Tasks:
- Utilize LSTM to forecast time series data.
- Implement ARIMA, GRU, and Prophet for comparative analysis and validation.
- Generate accurate predictions from our internal data.
Ideal Skills:
- Expertise in LSTM and other AI models.
- Proficiency in time series data analysis.
- Experience with ARIMA, GRU, and Prophet.
More about the project
I look forward to your bids. Please include your relevant experience and previous projects you have completed in this field.