A cost-valid tive Recommender System for taxi drivers (a recommended System for taxi drivers to effectively cost)

This is a 2014kdd, geographic location service paper.

Contents:

1) First, the author raised a valuable Question: recommend taxi drivers with routes that can be carried to the guests at the minimum cost.

2) Secondly, the novelty of the problem is that in the past, there were scattered recommendation points. This paper is a recommendation route. It can be said that the location service provided goes further.

3) then, two sub-problems are pointed out: a) how to calculate the benefits of each road section B) how to find the optimal road section from a complex road section

4) Next, the author lists the target functions for the first subproblem. For the second subproblem, the author proposes two methods: a) brute force method B) recursion method; at the same time, the author also puts forward K-top recommendations for overload situations to provide different routes for taxi drivers in the same location.

Inspiration:

1) raised the importance of a question: raised a valuable question is very important to the value of the paper

2) The solution to the problem in the paper is relatively not too advanced. Some tips can be used for reference when dealing with road data in the future.

3) in this paper, this problem is difficult to solve, but the author puts forward restrictions and simplifies step by step to make the problem easy to solve; for example, the time complexity of finding a route is n ** (M-1), but the author proves that when m is greater than 4, the yield growth rate is less than 10%, for taxi drivers, it is recommended that further road sections do not have much value, thus limiting m to less than 4, so the time complexity is barely acceptable.

4) find a suitable dataset. In this paper, the author does not have the road network information of San Francisco. Therefore, Google Maps are used to generate a map. Therefore, selecting a suitable lab location can reduce unnecessary work. In addition, the authors mentioned in this paper that the data they obtained contains the marks of taxi passengers and unloading passengers. This information is a decisive factor for them to solve the problem in this paper, without this data, it is difficult to identify whether the driver's journey is carrying passengers or roaming, and it is difficult to further calculate the route income. Therefore, finding the right data can solve the right problem.

"Paper download

A cost-valid tive Recommender System for taxi drivers (a recommended System for taxi drivers to effectively cost)