A carregar...
Publicação
Predicting Account Receivables Outcomes with Machine-Learning
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 7.07 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
The Account Receivables (AR) of a company are considered an important determinant of a company’s Cash Flow – the backbone of a company’s financial performance or health. It has been proved that by efficiently managing the money owed by customers for goods and services (AR), a company can avoid financial difficulties and even stabilize results in moments of extreme volatility. The aim of this project is to use machine-learning and data visualization techniques to predict invoice outcomes and provide useful information and a solution using analytics to the collection management team. Specifically, this project demonstrates how supervised learning models can classify with high accuracy whether a newly created invoice will be paid earlier, on-time or later than the contracted due date. It is also studied how to predict the magnitude of the delayed payments by classifying them into interesting, delayed categories for the business: up to 1 month late, from 1 to 3 months late and delayed for more than 3 months. The developed models use real-life data from a multinational company in the manufacturing and automation industries and can predict payments with higher accuracy than the baseline achieved by the business.
Descrição
Project Work presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence
Palavras-chave
Cash Flow Account Receivables Machine-Learning LightGBM