Binary or Ordinal Classification - Machine Learning Engineers & Research Scientists
Budget: $1,500 – $3,000 USD
Competent machine learning engineer research to attempt to solve a binary classification problem using a variety of techniques.
Specification of the problem is as follows:
- Training data is a few thousand rows by a few hundred columns (many redundant)
- Test data is also a few thousand rows
- Validation set will be withheld, and will consist of the most recent 30% of the available data.
- Target can be a choice of either a binary output, or alternatively an ordinal output.
Quite sparse in the training labels - approximately 1-10% positive label rate.
Any applicants will be interviewed before acceptance. If you are not prepared to have a 30 minute - 1 hour meeting beforehand then please do not apply or message.
Payment will be subject to satisfactory and meaningful outperformance of a baseline scikit learn ensemble model on the withheld validation set.
Project duration will be milestoned, and timeboxed to 14 days. There is the potential for future project work for outstanding cases.
In your bid, please indicate in some way your background and expertise, and demonstrate that you have appropriately understood the ask.
Specification of the problem is as follows:
- Training data is a few thousand rows by a few hundred columns (many redundant)
- Test data is also a few thousand rows
- Validation set will be withheld, and will consist of the most recent 30% of the available data.
- Target can be a choice of either a binary output, or alternatively an ordinal output.
Quite sparse in the training labels - approximately 1-10% positive label rate.
Any applicants will be interviewed before acceptance. If you are not prepared to have a 30 minute - 1 hour meeting beforehand then please do not apply or message.
Payment will be subject to satisfactory and meaningful outperformance of a baseline scikit learn ensemble model on the withheld validation set.
Project duration will be milestoned, and timeboxed to 14 days. There is the potential for future project work for outstanding cases.
In your bid, please indicate in some way your background and expertise, and demonstrate that you have appropriately understood the ask.