Skip to content

Latest commit

 

History

History
37 lines (18 loc) · 1.69 KB

EvaDB2.md

File metadata and controls

37 lines (18 loc) · 1.69 KB

Proposal2: Use evadb to predict network relationship or community

Introduction

In this project, I hope to use the evedb database and MySQL database for predicting friend relationships. We will store the user's relationship network and personal features in the MySQL database. Additionally, we will conduct relationship analysis using evedb, which includes tasks like link prediction or community classification.

System strategy

Using machine learning methods for graph analysis to predict relationships between nodes or communities. Storing user information in the database as user features while recording the network of relationships between users to conduct analysis on user relationships or communities.

Leveraging machine learning methods for relationship graph analysis.

Data is stored in a database, and AI research is conducted using evadb.

The aim is to provide predictive results, making it a reasonable backend application.

Challenges:

Machine learning predictions can be challenging, and the prediction process may take a considerable amount of time. There is a need to strike a balance between prediction time and accuracy.

The system may become quite complex, making it difficult to effectively utilize user features for predictions.

It may not be possible to convert the data into quantifiable values for analysis.

Use situation

User group analysis involves assessing the similarities among users and making similar recommendations to them.

Predicting communities leads to an enhanced user experience.

Efficient utilization of database content is aimed at making better use of database data.

Conclusion

I hope this is a meaningful project that can bring new functionality to evade.