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
Measuring and predicting social sustainability with machine learning: Creating social sustainability indices for municipalities in Sweden
KTH, School of Electrical Engineering and Computer Science (EECS).
2024 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesisAlternative title
Att mäta och förutspå social hållbarhet med maskininlärning : Ett framtagande av sociala hållbarhetsindex för kommuner i Sverige (Swedish)
Abstract [en]

Social sustainability is a large concept which has been measured multiple times with different data. However, these measurements gives a national measurement of how social sustainable a country is. This project creates three indices on a municipality level based on different weights, and compare these based on their ability to distinguish between municipalities, their predictability, and interpretability. The indices are created using Machine Learning, where the different weights are calculated using Principal Component Analysis. Further, future indices are predicted on a three- year basis to identify vulnerable municipalities. For this, Auto-regressive Integrated Moving Average (ARIMA), Support Vector Regression, and Random Forest are used. To evaluate the result the Mean Absolute Error (MAE) is calculated and the predicted index is transformed into labels showing an increase or a decrease after three years. The second principal component index effectively identifies vulnerable municipalities. For future predictions, the first principal component index in most cases achieves the best MAE, but the second principal component usually offers higher accuracy. Overall, the second principal component index stands out, though further data analysis is needed. Different principal components significantly impact index characteristics and model performance in measuring and predicting social sustainability.

Abstract [sv]

Social hållbarhet är ett omfattande begrepp som har mätts flera gånger med olika data. Dessa mätningar ger dock en nationell bedömning av hur socialt hållbart ett land är. Detta projekt skapar tre index på en kommunal nivå baserade på olika vikter, och jämför dessa utifrån deras förmåga att skilja mellan kommuner, deras förutsägbarhet och tolkbarhet. Indexen skapas med hjälp av maskininlärning, där vikterna beräknas med Principal Component Analysis (PCA). Vidare förutsägs framtida index på treårsbasis för att identifiera sårbara kommuner. För detta används Auto-regressive Integrated Moving Average (ARIMA), Support Vector Regression och Random Forest. Dessa modeller utvärderas med både Mean Absolute Error (MAE) samt med nogrannget där det förutsagda indexet omvandlas till etiketter som visar en ökning eller minskning efter tre år. Det andra indexet som är baserat på vikter från den andra principalkomponenten identifierar effektivt sårbara kommuner. För framtida förutsägelser uppnår det första indexet, baserat på den första principalkomponenten, den bästa MAE i de flesta fall. Det andra indexet, däremot, erbjuder högre noggrannhet i fler fall än de andra indexen. Sammantaget utmärker sig det andra indexet, baserat på den andra principalkomponenten, även om ytterligare dataanalys behövs. Olika principalkomponenter påverkar väsentligt indexens egenskaper och modellernas prestanda i att mäta och förutsäga social hållbarhet.

Place, publisher, year, edition, pages
2024. , p. 80
Series
TRITA-EECS-EX ; 2024:737
Keywords [en]
Social Sustainability, Machine Learning, Principal Component Analysis, ARIMA, Support Vector Regression, Random Forest
Keywords [sv]
Social Hållbarhet, Maskininlärning, Principalkomponentanalys, ARIMA, Support Vector Regression, Random Forest
National Category
Computer and Information Sciences
Identifiers
URN: urn:nbn:se:kth:diva-360095OAI: oai:DiVA.org:kth-360095DiVA, id: diva2:1938347
External cooperation
Action for Society
Supervisors
Examiners
Available from: 2025-02-20 Created: 2025-02-18 Last updated: 2025-02-20Bibliographically approved

Open Access in DiVA

fulltext(3570 kB)68 downloads
File information
File name FULLTEXT01.pdfFile size 3570 kBChecksum SHA-512
f1964b316f4b3557b09dd3d5b33a9183c4fdf89eb05179b0dbf2c443cfd2ad924b18fa16863920800e91fb8e5a8319b3ae7c1ee91bd7ca10769a23aab7cd6c8b
Type fulltextMimetype application/pdf

By organisation
School of Electrical Engineering and Computer Science (EECS)
Computer and Information Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 68 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: 265 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