Here are a few demos of the kind of problems which may come up in industry and which could be solved with data science and machine learning. These are small models and not intended for production use.
- Author prediction demo – a demo of a Convolutional Neural Network used for Forensic Stylometry. It is able to identify who wrote a document out of a set of known authors by constructing a ‘fingerprint’.
- Machine learning model predicting vessel detentions – a model in Microsoft Azure ML predicting which vessels in the port of Singapore are likely to be detained by port authorities, using live data feed of ships’ movements.
- Customer churn demo – a machine learning model predicting the likelihood of a customer ending a utility contract.
- Motion classification demo – a model that classifies accelerometer readings into actions such as walking, falling or sitting.