kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Logistic regression analysis of If’s car insurance market share: Purposeful selection of covariates benchmarked with elastic net, PCA and random forest
KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.).
2021 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

This project aims to build a logistic regression model of the car insurance market share of If Skadeförsäkring AB. The model is constructed by producing a set of candidates with reduced multicollinearity, subjecting each to purposeful selection of covariates and comparing the resulting summary statistics. These include the Hosmer-Lemeshow and Standardized Pearson goodness-of-fit statistics, pseudo-R^2, AIC, Mallows's Cp and AUC-ROC. The final model is examined by residual and influence diagnostics analysis. Each covariate is analysed in terms of response class frequency distribution, estimation impact and estimation impact over time. The model is benchmarked by simplistic applications of elastic net, PCA-preprocessed logistic regression and random forest. Throughout, a model trained on SMOTE oversampled data is fitted in parallel to investigate the effect of class imbalance. The final model performs in line (AUC ≈ 0.7) with the benchmark correspondents. The SMOTE versions underperform and overfit.

Place, publisher, year, edition, pages
2021.
Series
TRITA-SCI-GRU ; 2021:176
National Category
Mathematics
Identifiers
URN: urn:nbn:se:kth:diva-322424OAI: oai:DiVA.org:kth-322424DiVA, id: diva2:1719348
External cooperation
If Skadeförsäkring AB
Educational program
Master of Science in Engineering -Engineering Physics
Supervisors
Examiners
Available from: 2022-12-15 Created: 2022-12-15 Last updated: 2024-09-25Bibliographically approved

Open Access in DiVA

fulltext(2690 kB)532 downloads
File information
File name FULLTEXT01.pdfFile size 2690 kBChecksum SHA-512
a86cf516cce17fab7747295c0bcbaee7a6863e03191f31a204649c49833369427b3886f9c77025560c07a15a7f9a9f19c365504664d473852eded162cc119cdd
Type fulltextMimetype application/pdf

Authority records

Harting, Alice

Search in DiVA

By author/editor
Harting, Alice
By organisation
Mathematics (Div.)
Mathematics

Search outside of DiVA

GoogleGoogle Scholar
Total: 533 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 538 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf