An Empirical Study of Segmented Linear Regression Search in LevelDB

Round 1

Reviewer 1 Report

The manuscript titled An Empirical Study of Segmented Linear Regression Search in LevelDB scrutinizes a novel and scientific topic. The entire paper has been structured adequately notwithstanding the abstract must be developed still. 

The abstract should comprise the overall purpose of the manuscript, how to collect the raw dataset, and the new method that was applied by the paper. Of course, the findings should also be summarized there. 

If the abstract will be revised in this way, the manuscript may be accepted. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

it is an interesting work that proposes a learning model used to improve the search for information in large datasets, using OSM as an example, the 6 references that are over 5 years old must be modified to accept the paper.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

In this paper the authors use the linearity property of big data (keys and locations) and learned index concept to propose a novel search technique employing segmented linear regression. The techniques are implemented in LevelDB, which is one key-value store database. The two component or sub-techniques (equal-size and error-aware) are evaluated in terms of accuracy and segmentation overhead.  

The concept of learned index is presented briefly that introduces novel readers with this ML-approach for index management. Similarly, LevelDB structure is presented clearly that plays an important role to understand the linearity concept used in the storage and proposed technique.

The algorithm design, experiments and evaluation results are clearly described. Overall, the research goals are met, and the paper is complete. The only question is the suitability of this paper in Electronics journal. It seems there are many more suitable MDPI journals for this paper theme.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf