Wequity
- #Esg
Platform allows clustering algorithms, deployed on AWS server, clustered data (records with corresponding labels)
- Machine Learning
Impact
The client needed a solution for clustering articles and news from the past year for four different companies with varying amounts of documents.
Services we provided
Clustering algorithms
Deployed on AWS server, clustered data (records with corresponding labels)
Tech Stack
Python
Pytorch
Bert
HDBSCAN
UMAP
LDA
BertTopic
River
Challenges and Solutions
🧐 Challenges
- Identify compelling methods to extract relevant features.
- Select effective algorithms for grouping similar data points.
- Find the best feature extraction and clustering techniques for desired results.
💡 Solutions
Our developed solution encompasses the following steps:
- Data preprocessing and feature extraction. • Training, evaluating, and optimising hyperparameters for diverse clustering models.
- Deploying the solution on AWS, complemented by a user interface (UI) and an application programming interface (API).
User flow
1. User upload a database with news about different companies from the last year.
2. The system processes the input and generates insightful analytics regarding covered topics.