College Research
34 - Updates From Sir Till Now
Research Article Reviews¶
- Research Paper on Basic Parallel Processing
- A Bridging Model for Parallel Computation
- Drug Repositioning
- Drug Matrix
- Drug repositioning based on bounded nuclear norm regularization
- DRPADC A novel drug repositioning algorithm predicting adaptive drugs for COVID 19
- Drug Repositioning Concept, Classification, Methodology, and Importance in Rare Orphans and Neglected Diseases
Meeting Notes¶
Code Part¶
- Algorithm Docs
-
All equations must be numbered
- Figures should have caption below
- Table should have caption above (NO TABLE)
- Check if all references are cited
- Check the dpi for diagram 3 and 4 with DPI 300 or plus
- Add Result and Discussion
-
Remove abbreviations table
-
Call Sanika on Saturday for guiding on helping with the paper.
-
Images increase DPI for diagrams
- Decrease figure 1 and 2, 3 and 4 size to
- Add Graph and comparison points in result and discussion section
- Conclusion and Future Scope after results and discussion
-
Department of Computer Science and Engineering
-
From different co-ordinates of users sitting in different places
-
AUC or ROC for range 20 to 80.
- what would be better here for comparison
- Result representaiton curves
- - ROC is not suitable since this is not a classification task.
-
Add Suraj Sawant sir in authors
- In maps, add second comparative map route as well
--- AREA UNDER CURVE
3. How to Generate AUC/PR Curve for Your Case?¶
To compute AUC in your case, follow these steps:
Step 1: Define the Ground Truth¶
- Take your dataset of locations with their normalized density values.
- Define a binary relevance label:
- Relevant (1): If the location falls in the density range 20-80 and is near the user.
- Not Relevant (0): If the location falls outside the range or is farther away.
Step 2: Generate a Ranked List¶
- When a user queries for recommendations, your system returns a ranked list of locations based on density and proximity.
- Each recommended place is labeled as "Relevant (1)" or "Not Relevant (0)" based on density preference.
Step 3: Calculate Precision and Recall¶
- Compute precision (how many recommended places are relevant) and recall (how many relevant places are actually recommended).
- Vary a threshold on ranking (e.g., Top-5, Top-10, Top-15 recommendations) to calculate precision-recall values.
Step 4: Plot AUC-PR Curve¶
- The x-axis will be recall, and the y-axis will be precision.
- Compute the area under this Precision-Recall curve (AUC-PR) to summarize recommendation performance.
Breaking Down the Equation
- wdist → Weight for distance
- wdenw_{den}wden → Weight for density
- Haversine distance → Computes how far the place is from the user (in km or miles)
- \text{max_distance} → Maximum distance among all places (used for normalization)
- Density difference → Measures how much the place’s density deviates from the user’s preference
- \text{max_density_diff} → Maximum density difference in the dataset (used for normalization
Calculating the AUC curve using Recall and precision¶
- In order to plot the curve, we needed to set what a good recall value was.
- For this we set +- 20 from preferred density as an acceptable recall range, as every outcome is anyways going to be part of the user's density preference range.
-
Here what we will do is ask for recommendations and see how many of the recommended places actually fall in good range/ acceptable range.¶
Corrections for paper¶
- Tense should be present perfect until referring to something very old paper
- Instead of using "This Paper" say This study or This Work
- Redo citation no 29. it shouldn't be the part which defines what sections does what
-
%% This should be section C in literature Review%%
-
Pahile Flow diagrams mag output images
- How Our Proposed Solution Will Enhance Previous solutions will help madhe describe flow
- Detailed Structure of implementation, move it to the start of result and discussion
- Karan it is experimental setup
- R&D -> Detailed Structure of Implementation)
-
Move 1 & 2 to Result and Discussion
-
Move flow charts to proposed methodology and describe in PM.
R & D - Experimental Part - Dataset description - Performance Matrix Measures - Move equation 1 to 3 to Results and Discussions. - 1,2 and 5,6, in that context write what the result is. - Comparisons of Implementations
Project Report¶
- Intro
- Literature Review
- Research Gap and Problem Statement
- Research Gap
- Problem Statement and Objectives
- Proposed Methodology 1 / 50m - Title is first paper that has been accepted
- Proposed Methodology 2 / 50m -- Title for second paper for that has been accepted
- Result and Discussion
- Datasset
- Two sub points for two chapters %% %%
- experimental setup
- Sub points for different papers %% Already present as an individual chapter %%
- Perfromance Metrics
- Results and inferences
- Datasset
- Conclusion and Future Scope
-
Bibilography ( not a chapter references/bib at end directly)
-
Presentations would be in the same context, flow.
-
Update the PPT similarly as per this very flow. Update Beamer (PPT) to ESE
-
Abstract -> Remove Dean wali line, and add Dr Amit Joshi sir and Dr Suraj Sawant sir for the academic support and academic guidance.
-
Timeline July to May. Dolyana disel asa kara.
-[X] Recommendation output move text above the images -[X] Redo citations to have space before every citation 4.5 heading update to comparison 4.1 heading update to Experimental setup 4.1 last paragraph replace full stop with
3.1 Research Gaps 3.2 Problem statement and Objectives Roman numerals for first nonrelated pages
Experimental setup goes in results and discussion Dataset and data preparation If Experimental setup common make it a subsction in this chapter not a subsubsection
Dataset Experimental setup Performance Metrics 6.4. 1st paper title 6.5 2nd paper title
Conclusion and Future Scope
Appendix A Publication - 1st Accepted - 2nd Communicated Appendix B Timeline
Send Registration and acceptance mail to Joshi sir and Soma Ghosh ma'am