Classifying hot water chemistry: Application of multivariate statistics
0 likes589 views
The document discusses the classification of hot water chemistry using multivariate statistics, highlighting the importance of classifying geothermal systems. It details the methods employed, including hydrochemistry, multivariate analysis, and the use of R programming, while presenting results and preliminary remarks on data structure challenges. The closing remarks suggest further work for model validation and improvements in analysis techniques.
![Classifying hot water chemistry:
Application of multivariate statistics
Prihadi Sumintadireja[1], Dasapta Erwin Irawan*[1], Yuano Rezky[2], Prana Ugiana Gio[3],
Anggita Agustin[1]
[1] Faculty of Earth Sciences and Technology, Institut Teknologi Bandung,
Jalan Ganesa No. 10, Bandung 40132
[2] Ministry of Energy and Mineral Resources
[3] Faculty of Math and Natural Sciences, Universitas Sumatera Utara,
Jl. dr. T. Mansur No. 9, Medan 20155
* email: dasaptaerwin[@]outlook[.]co[.]id
* twitter: @dasaptaerwin ORCID (0000-0002-1526-0863)
1](https://image.slidesharecdn.com/20160331iigwmultivariat-160402222527/85/Classifying-hot-water-chemistry-Application-of-multivariate-statistics-1-320.jpg)
![Thank you
Classifying hot water chemistry:
Application of multivariate statistics
for more communications with the authors please:
- send your email to dasaptaerwin[@]outlook[.]co[.]id
- mention on twitter: @dasaptaerwin
- connect ORCID (0000-0002-1526-0863)
Dataset and R code are stored in RG:
http://dx.doi.org/10.13140/RG.2.1.3510.1205
26](https://image.slidesharecdn.com/20160331iigwmultivariat-160402222527/85/Classifying-hot-water-chemistry-Application-of-multivariate-statistics-26-320.jpg)
Classifying hot water chemistry: Application of multivariate statistics
- 1. Classifying hot water chemistry: Application of multivariate statistics Prihadi Sumintadireja[1], Dasapta Erwin Irawan*[1], Yuano Rezky[2], Prana Ugiana Gio[3], Anggita Agustin[1] [1] Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Jalan Ganesa No. 10, Bandung 40132 [2] Ministry of Energy and Mineral Resources [3] Faculty of Math and Natural Sciences, Universitas Sumatera Utara, Jl. dr. T. Mansur No. 9, Medan 20155 * email: dasaptaerwin[@]outlook[.]co[.]id * twitter: @dasaptaerwin ORCID (0000-0002-1526-0863) 1
- 2. versioning The following slides use versioning system on “Keynote”. This version has some improvements based on the presentation and some discussions 2
- 3. license I claim no right other than attribution 3
- 4. Background Classification is very important 4
- 5. Background Classification leads to characterisation of geothermal system 5
- 7. Background Classification: Direct method (geology, drilling, etc) → strongly qualitative Indirect method 7
- 8. Background Classification: Direct method (geology, drilling, etc) → strongly qualitative Indirect method → supports direct method, more quantitative 8
- 9. Method Here we used: Hydrochemistry, Multivariate analysis, R programming 9
- 10. Method Here we used: Hydrochemistry (relatively cheap), Multivariate analysis, R programming 10
- 11. Method Here we used: Hydrochemistry (relatively cheap), Multivariate analysis (robust and powerful), R programming 11
- 12. Method Here we used: Hydrochemistry (relatively cheap), Multivariate analysis (robust and powerful), R programming (robust, powerful, and free) 12
- 13. Method Samples = 11 Locations = Gorontalo Sources = EBTKE Dataset and code = available at zenodo.org (free repo) 13
- 14. Method 14 We combine: linear regression from multi-regression technique with axis rotation from PCA and CA.
- 15. Method 15
- 17. Results and discussions 17 Variables (elements) extracted from the dataset. We detect some strong correlations among the following elements.
- 19. Results and discussions 19 Our preliminary remarks is that we have three system running in the area (see the dots).
- 22. Closing remarks Regression tree technique has failed to read the data structure. Collinearity effect is major problem 22
- 23. Closing remarks PCA and cluster analysis have successfully classify the samples. Need more samples to validate training model. Hopefully we can make a useable model to classify other samples. 23
- 24. Closing remarks Proposed work around: manual variable selection then re-run the regression use other technique without regression principles, in this case: PCA and CA 24
- 25. Acknowledgment ITB for funding this research under 2016 ITB Research Grant Scheme. 25
- 26. Thank you Classifying hot water chemistry: Application of multivariate statistics for more communications with the authors please: - send your email to dasaptaerwin[@]outlook[.]co[.]id - mention on twitter: @dasaptaerwin - connect ORCID (0000-0002-1526-0863) Dataset and R code are stored in RG: http://dx.doi.org/10.13140/RG.2.1.3510.1205 26
- 27. Thank you R for all @dasaptaerwin 27