DATA MINING LECTURE 6 Similarity and Distance Sketching, Locality Sensitive Hashing.
Mining User Similarity Based on Routine Activities
-
date post
02-Feb-2016 -
Category
Documents
-
view
246 -
download
5
description
Transcript of Mining User Similarity Based on Routine Activities
kdd912@nchu 1
Mining user similarity based on routine activitiesPublished in:· JournalInformation Sciences: an International Journal archive
Volume 236, July, 2013
Pages 17-32 Elsevier Science Inc. New York, NY, USA Report:蔡宜真
Professor:Hsiao-Ping Tsai
kdd912@nchu 2
OutlineIntroductionSystem overview
Preliminary
ArchitectureRoutine activity mining Reference places extraction
Routine activities mining
User similarity calculation Reference place similarity calculation
Routine activity similarity calculation
ExperimentsConclusions
kdd912@nchu 3
IntroductionMobile user similarity is significant for location-based social network
services. With the pervasiveness of location-acquisition technologies, research on measuring mobile user similarity based on their trajectories has attracted a lot of attention.
In this paper, we address the problem of mining users’ long-term activity similarity based on their trajectories.
It propose a two-stage approach. At the first stage, the notion of routine activity is proposed to capture users’long-term activity regularities. The routine activities of a user are extracted from his/her daily trajectories. At the second stage, user similarity is calculated hierarchically based on the extracted routine activities.
kdd912@nchu 4
The daily routine activities of a student Tom can be summarized in Fig. 1. Since routine activity reflects both the temporal and the spatial regularities of people’s daily lives, we take it as the basis to measure the long-term similarity between two users.
kdd912@nchu 5
System overview2.1. Preliminary
Our system for user similarity estimation is based on routine activity extracted from raw GPS data. First, we clarify some concepts and their data representation, including GPS point, GPS trajectory, visit point, reference place, 1-day activity and routine activity.
kdd912@nchu 6
kdd912@nchu 7
2.2. Architecture
kdd912@nchu 8
Routine activity mining3.1Reference places extraction
kdd912@nchu 9
kdd912@nchu 10
3.2Routine activities mining
kdd912@nchu 11
An ideal clustering result should maximize the intra-cluster similarity (i.e. the average similarity between pairs of 1-day activities in the same cluster) as well as minimize the inter-cluster similarity (i.e. the average similarity between pairs of 1-day activities of different clusters), thus we used a variant of the Dunn index
(3)
kdd912@nchu 12
User similarity calculation4.1Reference place similarity calculation
(4)
(5)
(6)
kdd912@nchu 13
4.2 Routine activity similarity calculation
Two routine activities A 1 and A 2 , and their corresponding reference place sets PS1 and PS2 , their similarity can be calculated based on Eq. (8), where P 1i ∈ PS1, P2k ∈ PS2 , OMS is the Optimal Matching Sequence of A1 and A2 , min(A1 ℮ij , A 2 ℮ kj ) is the common probability of reference ‧ ‧places i and k within the jth time span, T is the number of time spans.
(7)
(8)
kdd912@nchu 14
Experiments
kdd912@nchu 15
kdd912@nchu 16
Conclusions In this paper, we propose an approach to measure user similarity for LBSNs
based on GPS trajectory mining. The most important novelty of our user similarity measure approach is that it can capture the similarity of users’ long-term activity regularities. To achieve this goal, we propose a framework to extract the routine activities from users’ daily GPS trajectories,and calculate the similarity score between users based on their routine activities.
While our approach exploits the GPS trajectories, it is important to extend the routine activity mining framework to make it compatible with other indoor locating infrastructure, and apply our approach to both indoor and outdoor environments. Another important issue is to mine user similarity with sparse and incomplete trajectory data. We consider these as prom-ising future works.