Automation is the core belief of DeeplogiX.
Problem statement: Unstructured data and dynamic content structures. We have to identify some unique Features from dynamic content and HTML structures.
- We gathered the training data from our scripts
Classification is done based on the knowledge base we had collected so far
- Feature synonyms we identified from our knowledge base
- Regrouped the training data with these custom sets to generate the final training data and took 80% of it for training
- Now we are running the model training and optimisations for unstructured data
- We are also using levenshtein method on pipe the output further to build the confidence score for the feature
Model and Our custom algorithm has started giving us 50% more accurate and faster results than earlier. Optimisations are continued
NEXT IS NOW…