Estimating the distribution of anthropogenic asbestos using random forest algorithm
Authors
Date
2020Access restrictions
The author has not given permission to make the work publicly available electronically. Therefore the material can be read only at the archival workstation at Jyväskylä University Library (https://kirjasto.jyu.fi/collections/archival-workstation).
Copyright
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Tässä tutkimuksessa selvitettiin koneoppimisella tapahtuvan datakäsittelyn käyttökelpoisuutta ja sen tuottamia tuloksia asbestin alueellisesta jakautumisesta Suomessa. Analysoitava data perustui rakennusten valmistumisvuoteen, käyttötarkoitukseen, tunnettuihin käytössä olleisiin asbestituotteisiin ja näiden jakautumiseen kuntien välillä. Analyysi tehtiin random forest-algoritmilla, jonka todettiin oppivan hyvin ennustamaan aineistoa, jonka suureiden ja tutkittavan asian välillä on selkeitä korrelaatioita.
Algoritmi onnistui arvioimaan asbestin määrää eri kuntien välillä, ja määrien huomattiin korreloivan rakennusten valmistumisajan kanssa, etenkin 1960-70 luvuilla, jolloin sitä käytettiin myös eniten. Erilliset asuinrakennukset edustivat suurinta osaa datasta, eikä tämän vuoksi kaikista rakennustyypeistä löydetty korrelaatiota asbestin kunnallisen määrän kanssa.
Tutkimuksen suurimmat ongelmat olivat tarkan, tilastoidun datan poissaolo asbestin käyttömääristä ja -kohteista, mikä loi epävarmuustekijöitä tuloksiin. Tästä huolimatta data käsiteltiin ja arvioitiin tarkkojen rakennustietokantojen avulla.
...
In this study machine learning was used to determine its value for evaluating asbestos amounts in Finland based on structure type and age data, known asbestos-based products, and their distribution between municipalities. The results derived from random forest algorithm showed that with a large number of attributes, that have a high correlation with the examined question the random forest algorithm was able to determine estimations of asbestos use in each municipality.
Asbestos amounts were noticed to correlate with built area of all buildings dating from 1921 to 1989. There was significant bias towards the building age, since most of the asbestos was used during 1960-70. Since small residential buildings dominated the data other building types were not found as significant correlators when estimating municipal-scale asbestos use.
The main problems of this study were the absence of accurate statistics of asbestos use in construction materials, which created uncertainties with the data. This uncertainty was accepted and data was imputed and evaluated with accurate statistical building data.
...
Keywords
Metadata
Show full item recordCollections
- Pro gradu -tutkielmat [29561]
Related items
Showing items with similar title or keywords.
-
Do Randomized Algorithms Improve the Efficiency of Minimal Learning Machine?
Linja, Joakim; Hämäläinen, Joonas; Nieminen, Paavo; Kärkkäinen, Tommi (MDPI AG, 2020)Minimal Learning Machine (MLM) is a recently popularized supervised learning method, which is composed of distance-regression and multilateration steps. The computational complexity of MLM is dominated by the solution of ... -
How can algorithms help in segmenting users and customers? : A systematic review and research agenda for algorithmic customer segmentation
Salminen, Joni; Mustak, Mekhail; Sufyan, Muhammad; Jansen, Bernard J. (Palgrave Macmillan, 2023)What algorithm to choose for customer segmentation? Should you use one algorithm or many? How many customer segments should you create? How to evaluate the results? In this research, we carry out a systematic literature ... -
Comparing the forecasting performance of logistic regression and random forest models in criminal recidivism
Aaltonen, Olli-Pekka (2016)Rikosseuraamusalalla on viime vuosina kehitetty uusintarikollisuutta ennustavia malleja (Tyni, 2015), jotka perustuvat tyypillisesti rekisteripohjaisiin mittareihin, jotka mittaavat mm. tuomitun sukupuolta, ikää, rikostaustaa ... -
Algorithmic issues in computational intelligence optimization : from design to implementation, from implementation to design
Caraffini, Fabio (University of Jyväskylä, 2016)The vertiginous technological growth of the last decades has generated a variety of powerful and complex systems. By embedding within modern hardware devices sophisticated software, they allow the solution of complicated ... -
Multilayer perceptron training with multiobjective memetic optimization
Nieminen, Paavo (University of Jyväskylä, 2016)Machine learning tasks usually come with several mutually conflicting objectives. One example is the simplicity of the learning device contrasted with the accuracy of its performance after learning. Another common example ...