Ranking Information Extracted from Uncertainty Quantification of the Prediction of a Deep Learning Model on Medical Time Series Data

Round 1

Reviewer 1 Report

The authors did great and thorough work based on their previous work to improve on classification focusing on biomedical time series data. 

Great work, and easy to read the manuscript. 

 

Few comments:

1. The authors used a lot of acronyms repeatedly, it affected the reading of teh manuscript. They have teh space and can use teh full expression except for well-known ones.
2. one acronym like the DNN model line 130. I am not sure what they refer to here?
3.  IN the "Introduction" section, I believe they need to support some statements with references like Lines: 16-17.
4. Figure 1: does not show teh axis units ( even though it seems obvious, but is recommended).

Author Response

I attach the reply to reviewers and the annotated document.

Author Response File: Author Response.pdf

Reviewer 2 Report

Paper presents a novel method for ranking information of uncertainty for medical time series using deep learning. The uncertainty at the higher level of several records for one patient is considered. The validation reports promising results.

Minor comments:

I advise considering using standard keywords to better present the research. Please revise the abstract according to the journal guideline. The abstract must be more informative on the method. elaborate on the DL method, please.

Further minor comments on developing the introduction section:

Although the paper has appropriate length and informative content, several parts must be improved and written in better grammar and syntax. It would be essential if authors would consider revising the organization and composition of the introduction, in terms of the definition/justification of the contribution/objectives, description of the former methods, the accomplishment of the former studies etc. Some parts of the introduction are generally difficult to follow. The paragraphs and sentences must follow the story. Kindly aim at elaborating on the concept, research gap, contribution, and the organization of the paper.

The motivation/research impact/application has the potential to be more elaborated. You may add materials on why doing this research is essential, and what this article would add to the current knowledge, etc. The originality of the paper can be discussed more. The research question must be clearly given in the introduction, in addition to some words on the testable hypothesis. Please elaborate on the importance of this work. Please discuss if the paper suitable for broad international interest and applications or better suited for the local application? Elaborate and discuss this in the introduction.

State of the art needs improvement. A detailed description of the cited references is essential. Several recently published papers are not included in the review section. In fact, the acknowledgment of the past related work by others, in the reference list, is not sufficient. Consequently, the contribution of the paper will be more clear. Furthermore, consider elaborating on the suitability of the paper and relevance to the journal. Kindly note that references cited must be up to date.

Comments on the presentation:

The quality and clarity of the figures must be improved. Figures 1, 3, 4, 5, 7, and 8 must be enlarged. currently, the data is not visible.

Comments on the method and results:  

Elaborate on the method used and why choosing this method. I suggest elaborating more on the DL used. Adding a brief description of the CNN-LSTM and discussing why it has been chosen can improve the methodology. 

Limitations can be also discussed. The research question and hypothesis must be answered and discussed clearly in the discussion and conclusions. Please communicate the future research; what method can be used in the future to improve the results?. The lessons learned must be further elaborated in the conclusion by discussing the results to the community and the future impacts. What is your perspective on future research?   

 

Author Response

I attach the reply to reviewers and the annotated document.

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper suggests that we can use uncertainty estimates in deep networks to improve classification accuracy, and turns to a particular dataset on saccades to establish this idea.

I literally have no critiques.  It's a bit trippy to think that we should send uncertainties to a random forest (along with means) to properly get a single label (with no error bars!) but it seems to work.

Author Response

I attach the reply to reviewers and the annotated document.

Author Response File: Author Response.pdf