search

Crispae / AOPbotvisualizer

Network analysis platform for evidence gathered using AOP-bot
https://ao-pbotvisualizer.vercel.app
2 stars 0 forks source link

Implement Filter options #6

Open Crispae opened 1 year ago

Crispae commented 1 year ago

Implement filter options to select only

  1. Atleast two types of nodes must be selected
  2. Relation type feature implement later
  3. Scoring filter
Crispae commented 1 year ago

Statistical tests to evaluate the significance of co-occurrence relationships.

  1. Chi-squared test: Determine if the observed co-occurrence is significantly different from what would be expected by chance.
  2. Fisher's exact test
  3. Pointwise Mutual Information (PMI): Measures the extent to which the co-occurrence of entities is more likely than random chance.
  4. Log-likelihood ratio: Identifies associations between entities while considering the expected frequency.

Apply the chosen statistical tests to the co-occurrence matrix to calculate p-values or other relevant statistics for each pair of entities.

Correct for multiple hypothesis testing using methods like Bonferroni correction or False Discovery Rate (FDR) correction to control for the family-wise error rate.

Crispae commented 1 year ago

Pointing towards some scoring, we can implement for co-occurrence scoring

  1. https://www.biorxiv.org/content/10.1101/444398v1
  2. https://aclanthology.org/D11-1098.pdf
  3. http://nlp.uned.es/~juaner/papers/jmromo18-aim.pdf
Crispae commented 1 year ago

https://doi.org/10.1016/j.jbi.2022.104047 https://www.science.org/doi/10.1126/sciadv.abj9204 https://link.springer.com/article/10.1007/BF00114885