I have got some categories that I need to classify my data into, I have already classified my data manually and I want to use this data to train my data. The description of data to be classified keeps on changing for example Land tax could be shown as
Statutory cost - Land Tax
Indirect Expense - Land Tax
Property - Land Tax
I won’t mind working on the classifying my data manually to train it but as I am newbie I don’t know where or which module to start this from? Any help on this would be appreciated.
Data to be classified
Manually Classified data
Book1.xlsx (11.3 KB)
if I got it right you are classifying costs and each cost has associated number in description. So if each number (talking about the first one) should always belong to same category why not simply write rules which will classify it. For it I would use Rule Engine (Dictionary) node as you possibly can have many rules.
Example rules from Rule Engine:
$column1$ LIKE "41000*" => "Rental Income"
$column1$ LIKE "42000*" => "Sundry rental expenses"
$column1$ LIKE "43000*" => "Other rental related income"
Thanks Ipazin, these numbers are 41000 or 42000 are not relevant as they will keep on changing or sometimes i might not even have number.
A sample of what I aim to classify is in the second image data to be classified.
Although I can use like function and match keywords but I sometimes it will give wrong results as same line have both keywords that is why I want to try and use the artificial intelligence model.
I would probably also take an approach similar to @ipazin. Maybe instead of numbers you could use key words, e.g. “Land Tax”, in the Rule Engine (Dictionary) node to classify your entries.
I think that using a machine learning model might be a bit out of proportion for your data. Since your data only contains two words sometimes, I think a simple Rule Engine approach would do the work. But if you really want to have look on how to classify text data using a machine learning model, you could have a look at this workflow for a start: Document classification.
Artificial intelligence model can give you wrong results as well
Thanks for guiding me @janina. I will try and look into it.
Ha ha that’s true. I am just trying to learn something new from something simple and then try to implement this learning on complex models
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.