A predictive model is only as good as the data used to build it. Many organizations are missing out on important opportunities to improve model performance because they’re ignoring important data – namely text data.
By some estimates, more than 80% of the data generated in organizations is in the form of free text data – less than 1% of which is effectively analyzed.
In this webinar, Ross Metusalem demonstrates how text data can be combined with non-text data to build better models and make better decisions. Register to watch Ross discuss:
- Potential applications from multiple industries of augmenting models with text data.
- How to mine text data for important features and summarize those features numerically so that they can be incorporated as factors in models.
- A case study that compares model performance with and without the variables obtained through text mining.