Data Classification for Financial Records

[Tenali Babu]

Hello,

I'm trying to use Simple ML's "Predict Missing Values" to classify financial transactions.  I have over 1000 records of golden data that has already been classified.  I'm following the instructions to use "Predict Missing Values" but the classification results are very poor.  I've tried to use ChatGPT plugins in Google Sheets which seem to work a lot better but very expensive.  I like that SimpleML is free but the results are very poor.  What can I do to improve my results? 

I've included a small sampling of the results below.  The column prefixed with "Pred:*" is generated by SimpleML. On the other other hand CategoryFull column contains  the correct categories that I would expect the model to predict.

Note that I do have a separate column named ComputedCategoryFull which is not shown in the screenshot below that i used for the "Column with Empty Cells".  This column contains about 2000 records of populated data and then another roughly 1000 records with empty data that needs to be predicted.

Any suggestions for what I could be doing to improve the classification results?