I have theoretical and programming knowledge regarding groundwater as well as Machine learning. Recently, I have studied many literature regarding application of machine learning in groundwater potential zoning. But I can not figure out, how can I incorporate this to find out groundwater potential zoning.
Dear Rashed Uz Zzaman,
the question is, what kind of data do you want to use and what do you want to analyze? If you want to analyze similar structures in pictures may be a Convolutional Neutral Network could be more helpful as Maschine learning. If the groundwater potential zoning are numbers with different coordinates, machine learning is better. For this problem, I can suggest you a support vector machine, which constructs a hyperplane or set of hyperplanes in a high- or infinite-dimensional space, which can be used for classification, regression, or other tasks like outliers detection.
I have seen you have been working on this work and you have some sequence of publications regarding the same problem statement (Underwater imagery).
Anyway, you are working on machine learning such as PCA, etc. There are different methods to apply machine learning for images. If your programming language is python I suggest you scikit-image( https://scikit-image.org/) or OpenCV(https://opencv.org/).
If you try to shift from machine learning to deep learning (As machine learning methods lack in processing high dimensional data effienctly) then you might use Keras or TensorFlow or PyTorch as your choice for building various kinds of neural networks (as explained by , @Geraldine Berger )
I hope it helps :)
Dear Rashed Uz Zzaman
First of all you need to have an approximate model for groundwater formation and this should be a data rich one. A typical generic model for groundwater formation should consider the:
1- Characterics of recharge and discharge zones.
2- The rain fall.
3- The topographic slope.
4- The characteristics of the drainage system in the study area.
5- The distribution of the natural vegetation.
6- The surface lithology of the area.
7- The structural geology
8- Existing wells and their patterns of spatial distribution.
In fact your machine learning model can be built around the characteristics of existing wells and how they are related to the other factors, which can be as features to train the model.
Dear Rashed Uz Zzaman
These articles might be useful have a look:
1. https://mdpi-res.com/d_attachment/water/water-13-00658/article_deploy/water-13-00658-v2.pdf
2. https://assets.researchsquare.com/files/rs-303246/v1_covered.pdf?c=1631860030
3. Article Using machine learning algorithms to map the groundwater rec...
4. Article Application of machine learning techniques in groundwater po...
5. https://www.nature.com/articles/s41598-021-85205-6
Kind Regards
Qamar Ul Islam
Qamar Ul Islam and Gamal Seedahmed Exactly .. I also read these articles and indicators. But I am confused how can they assign weight to every indicators with the help of machine learning.
because finally those weights are used in GIS to combine every indicators to delineate groundwater potential.
I have knowledge to assign weight using multi criteria analysis(MCA) which is more focusing of expert judgement. But now I want to convert my expertise regarding weight assign form MCA to ML.
I have works on this topic:
Safronova T. I. The problem of groundwater quality management in rice irrigation systems and the concept of its solution / T. I. Safronova, E. V. Lutsenko / / Polythematic network electronic scientific Journal of the Kuban State Agrarian University (Scientific Journal of KubGAU) [Electronic resource]. - Krasnodar: KubGAU, 2004. – №05(007). P. 116-124. - IDA [article ID]: 0070405011. - Access mode: http://ej.kubagro.ru/2004/05/pdf/11.pdf, 0.562 cu. p. l.
Safronova T. I. Synthesis, optimization and verification of the semantic information model of groundwater quality management on rice irrigation systems / T. I. Safronova, E. V. Lutsenko / / Polythematic network electronic scientific Journal of the Kuban State Agrarian University (Scientific Journal of KubGAU) [Electronic resource]. - Krasnodar: KubGAU, 2004. – №05(007). P. 125-132. - IDA [article ID]: 0070405012. - Access mode: http://ej.kubagro.ru/2004/05/pdf/12.pdf, 0.5 cu. p. l.
Safronova T. I. Cognitive structuring and formalization of the problem of groundwater quality management in rice irrigation systems / T. I. Safronova, E. V. Lutsenko / / Polythematic network electronic scientific Journal of Kuban State Agrarian University (Scientific Journal of KubGAU) [Electronic resource]. - Krasnodar: KubGAU, 2004. – №05(007). P. 133-147. - IDA [article ID]: 0070405013. - Access mode: http://ej.kubagro.ru/2004/05/pdf/13.pdf, 0.938 cu. p. l.
Safronova T. I. Investigation of the semantic information model of groundwater quality management on rice irrigation systems / T. I. Safronova, E. V. Lutsenko / / Polythematic network electronic scientific Journal of the Kuban State Agrarian University (Scientific Journal of KubGAU) [Electronic resource]. - Krasnodar: KubGAU, 2004. – №05(007). Pp. 148-171. - IDA [article ID]: 0070405014. - Access mode: http://ej.kubagro.ru/2004/05/pdf/14.pdf, 1.5 cu. p. l.
http://lc.kubagro.ru/aidos/The_Eidos_en.htm
Hi, Rashed
You can use the random forest (RF) model to predict the groundwater potential zones by incorporating the effects of different thematic layers (e.g., recharge, soil type, rainfall, drainage density, slope, etc., ).
- Badges
- Science topic