You are the owner of a restaurant. Recently, you are receiving high volumes of advance bookings for all tables that are placed at the restaurant. Despite receiving such high volumes of advance booking, you are not able to generate the targeted revenue. You suspect that the Chief Accountant is misreporting the total amount of bills that are generated.
To remove this issue, you require an automated system that can identify the total amount from a provided bill so that you can use it to check against the total earnings of your restaurant. Your task is to create a sophisticated NLP-based model that can process an invoice or receipt. Print the total amount mentioned in the processed invoice or receipt.
Data
- Train: Folder contains 190 PDF files carrying receipt details
- Test: Folder contains 70 PDF files carrying receipt details
- Train.csv: Contains training filenames with their corresponding bill total amount
- Test.csv: Contains test filenames
| Column name | Description |
| FileName | File name |
| Total Amount [Target] | Total amount extracted from corresponding bill/file |
Submission format
- You are required to upload your predictions in the .csv file [70 x 2 (excluding headers)] by clicking 'Upload File'.
- You are required to upload your source code or notebook file by clicking 'Upload Source File'.
- This is used by HackerEarth for offline evaluations to break ties if required.
Sample submission: Submission format expected
| FileName | Total Amount |
| Te_5 | 15.03 |
| Te_126 | 19.1 |
| . | . |
| . | . |
| Te_188 | 70.68 |
| Te_7 | 93.58 |
Evaluation criteria
score=100×r2_score(actual,predicted)
