Machine Learning
Generally, the features we used in the machine learning are:
We use luminosity intense to evaluate the street lamp condition in nighttime, which partly showed that if it is safe on road.
Vehicle information gathered by GPS, included speed and current coordinates are used as following: we sent coordinates to HERE API to get the on-road traffic jam factor. We can use this feature to infer the user’s mood.
Also, real-time weather information was used to get better recommendation result: we classify the weather conditions to three type: 1 for good weather, -1 for bad weather and 0 for modiocre weahther.
Furthermore, we divide a day to five period: morning, mid-day, off-work, night and midnight, and use time of the day as a machine learning feature. We believe that the people’s taste will change according to the time of the day.
- time of day
- luminosity intense
- weather
- traffic factor
- speed
We use luminosity intense to evaluate the street lamp condition in nighttime, which partly showed that if it is safe on road.
Vehicle information gathered by GPS, included speed and current coordinates are used as following: we sent coordinates to HERE API to get the on-road traffic jam factor. We can use this feature to infer the user’s mood.
Also, real-time weather information was used to get better recommendation result: we classify the weather conditions to three type: 1 for good weather, -1 for bad weather and 0 for modiocre weahther.
Furthermore, we divide a day to five period: morning, mid-day, off-work, night and midnight, and use time of the day as a machine learning feature. We believe that the people’s taste will change according to the time of the day.