Open Access

Identification and Analysis of Multi-tasking Product Information Search Sessions with Query Logs


Cite

Figure 1

Sample log records.
Sample log records.

Figure 2

Distribution of users by number of sessions.
Distribution of users by number of sessions.

Figure 3

Thresholds and F-scores of the clustering approaches.
Thresholds and F-scores of the clustering approaches.

Session with related search tasks.

SIDTimeQuery termsQuery terms (translation)
19852013/5/20 20:21:34休闲衬衫 男 短袖Casual shirt male short sleeve
19852013/5/20 20:22:10休闲衬衫 男 绿Casual shirt male green
19852013/5/20 20:23:43短裤 男Shorts male

Distribution of the sessions according to the number of task per session.

Number of task in a sessionFreq.Percent (%)Cumulative percent (%)
1214061.461.4
274821.582.9
32928.491.3
41323.895.1
5732.197.2
6371.198.2
7230.798.9
8140.499.3
9110.399.6
1050.199.8
11 and more80.2100

Three related search tasks.

SIDTimeQuery termsQuery terms (translation)
8792013-05-04 12:06:25增高鞋真皮休闲Hidden heel shoes leather leisure
8792013-05-04 12:20:25夏季潮男洞洞鞋牛皮Summer male leather crocs
8792013-05-04 13:05:58万斯低帮豹纹Vance leopard print low-cut

Average session duration.

Session typeAverage session duration
One task36 minutes 9 seconds
Two tasks54 minutes 19 seconds
Three or more tasks1 hour 22 minutes 22 seconds

Average query length.

Session typeAverage query length in characters
One task7.56
Two tasks7.28
Three or more tasks7.32

Sample of human task identifications.

SidQuery Terms (original)Query Terms (translation)Coder
#1#2
11973丰胸仪Breast augmentation instrument11
21973优格格丰乳仪Yougege breast augmentation instrument11
31975北京茶月饼Beijing tea mooncake22
41975金凤呈祥Jinfengchengxiang33
51975金凤呈祥 200Jinfengchengxiang 20033
61976美优食品Meiyou food44
71977XQB38-83皮带XQB38-83 belt55
81978味多美卡Meiduomei gift card63
91978Laver丰胸精油Laver breast augmentation oil76
101978AOC 拉莫圣日尔曼干红葡萄酒 750mlAOC Saint Germain Rameau claret 750ml87
111978AOC银奖圣玛杰庄园干红葡萄酒 750mlAOC silver award Domaine Saint Majan claret 750ml87
121978圣玛杰庄园干红葡萄酒 750mlDomaine Saint Majan claret 750ml87
131978红绳Red rope98
141978红绳批发Red rope wholesale98
151978项链挂绳编织Necklace rope woven98

Basic characteristics of sessions and tasks.

ItemBasic characteristics
Average number of queries per session2.57
Highest number of queries in a session21
Average number of tasks per session1.78
Highest number of tasks in a session41
Average number of queries per task1.45
Highest number of queries in a task15

Sample query records.

User IDSidQuery termsQuery terms (translation)
10284339747169671481973丰胸仪Breast augmentation instrument
10284339747169671481973优格格丰乳仪Yougege breast augmentation instrument
10284339747169671481975北京茶月饼Beijing tea mooncake
10284339747169671481975金凤呈祥Jinfengchengxiang
10284339747169671481975金凤呈祥200Jinfengchengxiang 200
10284339747169671481976美优食品Meiyou food
10284339747169671481977XQB38-83皮带XQB38-83 belt
10284339747169671481978味多美卡Meiduomei gift card
10284339747169671481978Laver丰胸精油Laver breast augmentation oil
10284339747169671481978AOC 拉莫圣日尔曼干红葡萄酒 750mlAOC Saint Germain Rameau claret 750ml
10284339747169671481978AOC银奖圣玛杰庄园干红葡萄酒 750mlAOC silver award Domaine Saint Majan claret 750ml
10284339747169671481978圣玛杰庄园干红葡萄酒 750mlDomaine Saint Majan claret 750ml
10284339747169671481978红绳Red rope
10284339747169671481978红绳批发Red rope wholesale
10284339747169671481978项链挂绳编织Necklace rope woven

Average number of queries per session and per task.

Session typeNumber of queries per sessionNumber of queries per task
One task1.451.45
Two tasks2.931.47
Three or more tasks6.141.43

Average duration of tasks.

Session typeAverage duration of tasks
One task37 minutes 10 seconds
Two tasks27 minutes 26 seconds
Three or more tasks19 minutes 44 seconds

Task identification results.

MethodDictionaryThresholdRPF
Rule basedDefault0.30.89950.85420.8763
Rule basedDefault0.350.86700.89460.8806
Rule basedDefault0.40.84140.92320.8804
Rule basedDefault + Pro-Catalog0.30.88370.85520.8692
Rule basedDefault + Pro-Catalog0.350.84530.89260.8683
Rule basedDefault + Pro-Catalog0.40.82660.90540.8642
Clustering-avgDefault0.30.83940.91920.8775
Clustering-avgDefault0.350.80790.93790.8681
Clustering-avgDefault0.40.77240.95270.8531
Clustering-avgDefault + Pro-Catalog0.30.82460.92320.8711
Clustering-avgDefault + Pro-Catalog0.350.78920.93790.8571
Clustering-avgDefault + Pro-Catalog0.40.75570.95370.8432
Clustering-maxDefault0.30.91230.83840.8738
Clustering-maxDefault0.350.88670.87780.8822

This approach yields the highest F-score and is used to perform task identification for the rest of the dataset.

Clustering-maxDefault0.40.85120.91230.8807
Clustering-maxDefault + Pro-Catalog0.30.90150.84330.8714
Clustering-maxDefault + Pro-Catalog0.350.86500.87880.8719
Clustering-maxDefault + Pro-Catalog0.40.81380.91430.8611

Three unrelated search tasks.

SIDTimeQuery termsQuery terms (translation)
135272013-05-29 18:27:16海狮油Sea lions oil
135272013-05-29 18:35:24上海移动100元快充Shanghai Mobile 100 yuan recharge
135272013-05-29 18:35:48上海移动10元Shanghai Mobile 10 yuan
135272013-05-29 18:35:58上海移动100元Shanghai Mobile 100 yuan
135272013-05-29 19:05:44澄糖滋润护唇膏玫瑰粉红Sugar moist lip balm rose pink

Session with unrelated search tasks.

SIDTimeQuery termsQuery terms (translation)
678042013-05-17 10:51:33内存卡16g正品包邮Memory card 16g free delivery
678042013-05-17 10:53:35vip16g正品包邮Vip memory card 16g free delivery
678042013-05-17 11:12:08水杯Water cup

Session duration.

ItemSession duration
Average session duration49 minutes 3 seconds
Average task duration27 minutes 36 seconds
Longest session14 hours 56 minutes 22 seconds
eISSN:
2543-683X
Language:
English
Publication timeframe:
4 times per year
Journal Subjects:
Computer Sciences, Information Technology, Project Management, Databases and Data Mining