Prediction of Prospecting Target Based on Selective Transfer Network
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
It is an interesting paper suggesting introducing artificial intelligence technology to conventional geological prospecting. Without hesitation, such technology employment significantly increases the effectiveness of the geological prospecting for ore and other economic minerals.
However, such an investigation should include much-extended background (in the Introduction). I suggest including the following sources:
Duda, R.O. and Hart, P.E., 1973. Pattern classification and scene analysis}, John Wiley & Sons, N.Y.
Eppelbaum, L., Eppelbaum, V. and Ben-Avraham, Z., 2003. Formalization and estimation of integrated geological investigations: Informational Approach. Geoinformatics, 14, No. 3, 233-240.
Some minor remarks:
4. It should be 'Experiments,' but not 'Experiment'.
Lines 340-341 (Conclusions): should be 'The main conclusions are:'
After the corresponding revision, this MS can be accepted for publication.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Prediction of prospecting target based on selective transfer network
Overall:
The work proposes a combination of techniques involving transfer learning and a modified kernel for convolutional neural networks using a massive deep architecture for the prediction of prospecting geological samples. The introduction and region of interests is well described with some recommendations. The methodology and computational approach is well written. The results show that the developed method outperforms state-of-the art methods but a more elaborated presentation of the results is required.
Mayor Corrections:
1- Introduction:
The motivation behind the project in lines 19-113 is well described. However, in order to make it more understandable it is recommended to establish, before Fig. 1, what is the shape and characteristics of the input and output vectors. Are they a sequence of composition or lithographic data for previous samples. What are they? This will give the reader a better idea of what you are explaining.
2 – Study area:
Line 131, what is 1:50,000 please explain.
Line 140: what is AUC
Line 179: How was the value of the variance selected?
As variables only account for x and y coordinates is it independent of the depth f the site? Please explain
3- Methodology:
Section 3.3:
Line 224: What about overfitting? Which is the ratio of the final number of weights to the number of data points?
4- Results:
Given the methodology involved and the volume of concepts developed in the manuscript the amount of presented results is scarce. The table presenting the evaluation parameters is clear. However, more visualization of the errors while predicting the prospecting of the mineral samples and the quality of the sample is needed. Please elaborate more the results in order to visualize the predicted results more clearly.
Minor Corrections:
Line 138 – 7,237 points
Comments for author File: Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors have updated the manuscript and presented an improved version. This reviewer finds the new manuscript in a better shape and believes that it is a valuable material for publication in the journal.