TaxiRec （阅读笔记）20180119

《TaxiRec: Recommending Road Clusters to Taxi Drivers Using Ranking-based Extreme Learning Machines》ACM SIGSPATIAL 2015

1、Abstract

Propose TaxiRec: a framework for discovering the passenger-finding potentials of road clusters, which is incorporated into a recommender system for taxi drivers to hunt passengers.（提出了一个出租车寻客推荐模型）

• In TaxiRec, we first construct the road network by defining the nodes and road segments. （根据nodes和road segement 构建路网）

• Then, the road network is divided into a number of road clusters through a clustering process on the mid points of the road segments.（将路网分割为一系列的clusters）

• a set of features for each road cluster is extracted from real-life data sets, and a ranking-based extreme learning machine (ELM) model is proposed to evaluate the passenger-finding potential of each road cluster.（为每个road cluster 提取特征，用ELM模型给这些cluster的上车潜能排序）

2、Highlight

It is noteworthy that existing recommender systems for taxi drivers focus on recommending the passenger-finding strategy in a certain region or the most efficient path to a given destination. Different from the existing works, TaxiRec treats each road cluster as an evaluation unit and recommends the most potential road clusters to taxi drivers.
翻译过来就是：
本系统和主流的推荐系统的区别：

• 主流的系统：
  推荐在某区域的寻客策略
  or 推荐一条去往给定终点的有效路径
• 本系统：
  把road cluster作为最小单元
  直接推荐road cluster

3、Main idea

The basic idea is to treat each cluster as a sample $(x_i, y_i)$, which is composed of n input features $x_i = [x_i1, . . . , x_in]^T (n = 15 here)$, and one output decision value yi (the number of pick-up actions). Then, a regression model could be constructed on labeled samples, in order to find the relationship between $x_i$ and $y_i$. This model is then used to predict the decision value of unlabeled samples.

可以看做是 一个线性方程的回归问题，输入自变量x，得到因变量y，相对于BP算法的好处是，本算法不需要迭代。

4、总结

感觉帮助不大