uncg-cse / bioinformatics_trends Goto Github PK

Implement Random Forest and XGB classification on the dataset and try different ways for dividing the classes for each.

Explore data on authors and how they might relate to citations. Produce at least 1 graph to illustrate your findings.

Deciding which ML algorithms are best suited for our project

Generate Probability Distribution Functions and Box plots to better understand the field of Author Keywords

This issue was started and completed.

Research NLP and how it relates to our project and goal. Produce specifics for the next team meeting concerning how or why we should use NLP in the context of our project.

Find the best approach to divide the rows into different classes for classification type ML

Merge Mouna, Logan and Darpan's files together and Merge Luke and Steve's files separately

This was research for whether this particular algorithm is useful or not to our project

Completed research to better understand how to use NLTK to prepare our non-numerical data for machine-learning processing.

authors
author id
sequence number
affiliation id
indexed-name (first name, middle initial, last initial)

Script preparation for the presentation

• Generate two bar graphs relating average number citations per publication to publication subject
• Generate box plots for the same graphical analysis.
• Compare probability distributions of different subjects' avg citations per publication.
• Compare two groups:

1. subjects w/ most overall number publications
2. subjects w/ most overall number citations

https://github.com/UNCG-CSE/Bioinformatics_Trends/blob/master/Data%20Exploration/Logan/Logan_p3.ipynb
https://github.com/UNCG-CSE/Bioinformatics_Trends/blob/master/Data%20Exploration/Logan/Logan_p3_boxplot.ipynb

Work on author keyword analysis and how the author keywords corresponds to number of citations

I created a probability distribution for the average number of citations per paper (grouped by subject area)

Create a correlation graph between length of title of a paper and number of citations

I summed the total citations for each subject area. Then divided total citations by number of papers to get the average citations per paper. Then created bar plot comparing each subject area.

Find out which ML Algorithms would work best for the available dataset.

Map is a representation of the number of publications of a respective country. In this case, I used the top ten.

Produce at least one graphical representation of the citation data in whatever forms you judge to be useful.

The is a generic issue created to show an example.
LDA - code to create keywords classifying as such.