I totally agree with you. I bought this app several years ago and during the time of its usage I needed this feature several times. Today I understood there's no good place for a new type of my expense. No existing category is suitable for me but I'm allowed only to create a sub-sub-category. And it's quite frustrating.
Moreover, I insert all my records manually so I believe the ML feature, which support mentions all the time, is not even in use for me, so your suggestion looks very suitable in my situation.
And as a software engineer myself I understand that most likely there IS a solution to this request even though there's some ML algorithm that relies on hardcoded data and it's a shame support just keeps declining these types of requests without considering the change more seriously.
Budgetbakers, here are a couple more thoughts on how you can implement it:
1. Give the option to create new sub-categories but show a warning that the ML feature/features will not work with them.
2. Collect a larger dataset to make your model recognize a wider set of categories.
3. Collect data for the dataset from the application itself. Users will manually assign correct categories to their expenses, so just keep metadata and train the model using it.
I totally agree with you. I bought this app several years ago and during the time of its usage I needed this feature several times. Today I understood there's no good place for a new type of my expense. No existing category is suitable for me but I'm allowed only to create a sub-sub-category. And it's quite frustrating.
Moreover, I insert all my records manually so I believe the ML feature, which support mentions all the time, is not even in use for me, so your suggestion looks very suitable in my situation.
And as a software engineer myself I understand that most likely there IS a solution to this request even though there's some ML algorithm that relies on hardcoded data and it's a shame support just keeps declining these types of requests without considering the change more seriously.
Budgetbakers, here are a couple more thoughts on how you can implement it:
1. Give the option to create new sub-categories but show a warning that the ML feature/features will not work with them.
2. Collect a larger dataset to make your model recognize a wider set of categories.
3. Collect data for the dataset from the application itself. Users will manually assign correct categories to their expenses, so just keep metadata and train the model using it.