Basics Of Research

By gaganyatri

Few Observations for Research

Dataset Download

While using any public dataset, collect information regarding : 
	1. Size of Dataset, 
	2. No. of paper citations ( Usage of Dataset) 
	3. State of the art algorithms,
	4. Constraints of the data
	5. Errors Identified with data

Literature Survey

Collect following information for chosen domain
	1. Top 10 - Conferences  - Also their acceptance criteria and Deadlines
	2. Top Journals for domain
	3. State of The Art Papers in the domain/ related
	4. Initial Paper/Discovery which led to the domain
	5. Challenges faced in each paper
	6. Dataset used for the paper
	7. Open Issues and Future work of paper

Calculations

accuracy = (correctly predicted class / total testing class) × 100%

The accuracy can be defined as the percentage of correctly classified instances (TP + TN)/(TP + TN + FP + FN). where TP, FN, FP and TN represent the number of true positives, false negatives, false positives and true negatives, respectively.
also you can use standard performance measures:

1. Sensitivity = TP / TP + FN
2. Specificity = TN / TN + FP
3. Precision = TP / TP + FP
4. True-Positive Rate = TP / TP + FN
5. False-Positive Rate = FP / FP + TN
6. True-Negative Rate = TN / TN + FP
7. False-Negative Rate = FN / FN + TP