Naver’s new travel recommendations are taking away the hassle of having to plan out your day while on holiday. The corporation is using its new AI-based context recognition technology ‘ConA’ to offer personalized travel itineraries and information.
The technology takes over 12.2 million reviews, blogs and other content from travelers over the past two years, and analyzes the big data for keywords. It then uses deep learning technology to link the keywords together and automatically create themed itineraries. Naver says that because ConA will continue to learn from new data, the it’s recommendations will become better as time goes on.
Users can browse from a list of tours with one or two main keywords, each one with its own route and list of destinations. Each tour includes information such as travel time and distance between locations, as well as an estimate of how long the whole itinerary would take. Reviews about each stop can also be viewed.
One interesting point is that the platform uses Google Maps to plot the generated route for each themed tour, but as Naver’s maps only cover Korea it doesn’t have much choice. Most routes are kept fairly short in distance, and from a quick browse through various cities, the longest ones were around 11-12km.
One frustrating issue is that the Google Maps route provided is for travel by car, while most travelers would be more likely to use public transport. You can switch to public transport mode, but you end up losing all but your first destination which isn’t much help.
For now ConA’s system is limited to Korean, but Naver says that it’s planning to apply the same technology to English, Chinese and Japanese content in the future.
