irthomasthomas
commented
8 months ago \n gh formatting\n\nBy default, the result of gh commands are output in line-based plain text format.\nSome commands support passing the --json flag, which converts the output to JSON format.\nOnce in JSON, the output can be further formatted according to a required formatting string by\nadding either the --jq or --template flag. This is useful for selecting a subset of data,\ncreating new data structures, displaying the data in a different format, or as input to another\ncommand line script.\n\nThe --json flag requires a comma separated list of fields to fetch. To view the possible JSON\nfield names for a command omit the string argument to the --json flag when you run the command.\nNote that you must pass the --json flag and field names to use the --jq or --template flags.\n\nThe --jq flag requires a string argument in jq query syntax, and will only print\nthose JSON values which match the query. jq queries can be used to select elements from an\narray, fields from an object, create a new array, and more. The jq utility does not need\nto be installed on the system to use this formatting directive. When connected to a terminal,\nthe output is automatically pretty-printed. To learn about jq query syntax, see:\nhttps://stedolan.github.io/jq/manual/v1.6/\n\nThe --template flag requires a string argument in Go template syntax, and will only print\nthose JSON values which match the query.\nIn addition to the Go template functions in the standard library, the following functions can be used\nwith this formatting directive:\n\n autocolor: like color, but only emits color to terminals\n color
[ ] python - using scikit-learn to classify multiple outputs of banking transactions - Stack Overflow
Quote
To understand better : do you already know the number and names of class ? Do you already know what can be all variable use for classification ? – Dadep Feb 15, 2017 at 10:19 Yes I know the number of classes and names as these are parsed from a ledger input file and the idea is to try to classify to a previously defined class (ledger destination account and payee). I think there is a typo in your second question...yes I know what can be available for use for classification. – Jeff M Feb 15, 2017 at 14:53 So you can use any method of multi class supervised machine learning for classification... you can start with something simple to understand like naive bayes (analyticsvidhya.com/blog/2015/09/naive-bayes-explained, and, scikit-learn.org/stable/modules/naive_bayes.html). But first of all you have to prepare your data, make it in format that your classifier would be able to use. I had a look on your code but I don't understand very well your data – Dadep Feb 15, 2017 at