I noticed that we can use Facebook Prophet for univariate and Multivariate type time-series datasets.
Univariate - Date, target (eg. Payments amount)
Multivariate - Date, target (eg. Payments amount), another related field (eg. Production units)
In my dataset, I have to identify anomalies in each vendor wise.
eg. Date, Vendor Name, Payment amount
Is it possible to use this type of dataset to identify anomalies through facebook prophet time series modeling?
Do I have to 1. slice the dataset into parts (eg. vendor-wise) and analyze it with Facebook Prophet separately for each vendor,
2. add a new column showing the bin value based on payment amount, use it with the multivariate method?
Is there any other better way/method to consider to identify anomalies?
Appreciate it if someone can help
Univariate - Date, target (eg. Payments amount)
Multivariate - Date, target (eg. Payments amount), another related field (eg. Production units)
In my dataset, I have to identify anomalies in each vendor wise.
eg. Date, Vendor Name, Payment amount
Is it possible to use this type of dataset to identify anomalies through facebook prophet time series modeling?
Do I have to 1. slice the dataset into parts (eg. vendor-wise) and analyze it with Facebook Prophet separately for each vendor,
2. add a new column showing the bin value based on payment amount, use it with the multivariate method?
Is there any other better way/method to consider to identify anomalies?
Appreciate it if someone can help