Skip to content

fintechees/SERP-Experiment-Engine

Repository files navigation

SERP Structure and Visibility-Weighted Perception Formation

A structural framework for studying how ranking visibility influences user perception in Search Engine Result Pages (SERPs).

This project investigates how the visibility distribution of controversy-oriented or negatively associated search results affects perceived acceptability of entity-oriented SERPs. Unlike traditional Information Retrieval studies that primarily focus on relevance ranking, this framework examines SERPs as perception-forming structures.

The study introduces:

  • A SERP Stimulus Tensor structure
  • Semantic severity manipulation
  • Positional visibility manipulation
  • Deterministic experimental assignment
  • Controlled SERP reordering methodology
  • Interactive perception evaluation experiments

The repository includes:

  • Research manuscript
  • Experimental methodology
  • Visualization dashboards
  • Reproducible experiment generator
  • Data analysis charts
  • Supplementary materials

Experimental Dashboard

Interactive visualization dashboard:

https://xxx.slander.ai/

The dashboard contains:

  • Semantic severity curves
  • Positional visibility curves
  • Experimental charts
  • Aggregated perception results

Repository Structure

slanderpilot.html              Stimulus Generator
serp-perception-research.pdf   Experimental Paper
semantic.png                   Experiment Results
position.png                   Experiment Results

Research Focus

This work explores the hypothesis that:

Visibility hierarchy within search engine results contributes significantly to entity perception formation.

The study focuses on structural visibility effects rather than content removal or censorship questions.


arXiv Submission Notice

This repository accompanies an upcoming arXiv submission.

As an independent researcher submitting to arXiv for the first time, endorsement is currently required for several relevant subject categories.

If you are eligible and willing to provide endorsement after reviewing the repository or manuscript, your assistance would be greatly appreciated.

Endorsement code:

Q7EWVP

Thank you for your time and consideration.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages