predictive effect in the classification made according to the comments in different fields
I want to do a classification through comments categorized in 4 areas(X,Y,Z,M). Categorizing the product as good or bad based on the comments in the fields X, Y, Z, M. How can I follow a path to see the effects of these 4 areas on the result.
For example;
| Id | X | Y | Z | M | Result |
|---|---|---|---|---|---|
| 1 | The prod.. | I fell.. | Very bad.. | lost of time.. | 0(bad) |
Using this data, the model will be given comments in the x, y, z, m fields and it will be provided to classify according to the score in the result table. Next, I want to find out in which the comments of area contributed the most to the model's conclusion. Example, Z.
It was first to combine the comments in the 4 fields, then to predict via lstm, but the separation of the fields is lost. Would it be a good solution to build and combine 4 different models?
Topic text-classification lstm deep-learning sentiment-analysis nlp
Category Data Science