Name
Abstract | ||
|---|---|---|
We identified two use cases for machine learning (ML) in the process of handling insurance claims related to water damage. The first use case is a classification task and concerns the decision, whether or not to send an expert to review the water damage. The second use case is a regression task and involves estimating the initial reserve to declare for the claim. We compared the performance of different ML models for both use cases. Concerning both the classification and regression task, a neural network (NN) with four layers and about 9,000 parameters performed the best. On the classification task, the NN has a precision of 0.8093. On the regression task, the NN has a mean absolute error (MAE) of 1,799 EUR and outperforms human estimations, which have an MAE of 2,258 EUR. |
Year | DOI | Venue |
|---|---|---|
2021 | 10.1007/978-981-16-2380-6_82 | PROCEEDINGS OF SIXTH INTERNATIONAL CONGRESS ON INFORMATION AND COMMUNICATION TECHNOLOGY (ICICT 2021), VOL 2 |
Keywords | DocType | Volume |
Machine learning, Insurance claims, Initial reserve, Classification, Regression | Conference | 236 |
ISSN | Citations | PageRank |
2367-3370 | 0 | 0.34 |
References | Authors | |
0 | 7 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Michael Dammann | 1 | 0 | 0.34 |
| Nicolai Gnoss | 2 | 0 | 0.34 |
| Pamela Kunert | 3 | 0 | 0.34 |
| Eike-Christian Ramcke | 4 | 0 | 0.34 |
| Tobias Schreier | 5 | 0 | 0.34 |
| Ulrike Steffens | 6 | 0 | 0.34 |
| Olaf Zukunft | 7 | 0 | 0.68 |