How to use ML to detect key from set of value(s)

Atul Sureka 1

I have a CSV file with first row as header & after that each row has set of values for each column. Each row in the CSV correspond to some field in the DB. user has to manually map the header attributes to Business fields (or DB fields).

I want to use ML to learn from existing values, when single or set of values are provided then it should detect the key. Overall I want to avoid manual mapping. One can think going forward there is no more header row in the CSV file.

Is any services (preferably Azure) to machine learn from set of key-value pairs. When a value is provided then it should detect the keys? Please note this does not involve any OCR, my data is in CSV form.

Atul

romungi-MSFT 41,961 Reputation points Microsoft Employee

2021-10-04T09:36:02.753+00:00

@Atul Sureka A service that is available to detect time series anomalies without deep understanding of Machine Learning is the Azure Anomaly detector. This can basically detect any outlays from time series data or metrics. The mutlivariate option might work for you, You can explore this option based on your use case.

If this option does not work then Azure Machine Learning helps you with a workspace and designer tool to create custom models using algorithms of your choice. This requires basic and intermediate understanding of working with ML models to train and infer from the experiments.