Natural Language Entity Clustering for Data Visualization

The patent introduces natural language processing as an intermediary layer between user input and statistical clustering. The NLP component processes and structures unstructured natural language text into identifiable entities and driving factors, which then feed into the clustering algorithm. This intermediary transformation enables accurate clustering from casual user input without requiring users to manually structure their queries or understand complex data formats.

The system performs preliminary action by proactively generating and asking follow-up questions to gather necessary driving factor values before performing the statistical clustering. Rather than waiting for users to provide all necessary information or performing broad analysis and then filtering results, the system identifies what information is needed and collects it in advance through targeted questions, ensuring the analysis is relevant and focused from the start.

The patent applies segmentation by dividing the large dataset into distinct clusters based on the identified driving factors and their values. Instead of visualizing all data points together, the system segments the data into meaningful groups (clusters) that share similar characteristics. This segmentation enables focused visualization of specific segments relevant to the user's query, providing precise insights without overwhelming the user with the entire dataset.