TY - JOUR
T1 - An intelligent framework for auto-filling web forms from different web applications
AU - Wang, Shaohua
AU - Zou, Ying
AU - Keivanloo, Iman
AU - Upahyaya, Bipin
AU - Ng, Joanna
N1 - Funding Information:
We would like to thank all the IBM researchers at IBM Toronto Laboratory CAS research for their valuable feedback on this research. This research is partially supported by IBM Canada Centres for Advance Studies.
Publisher Copyright:
© 2017 Inderscience Enterprises Ltd.
PY - 2017
Y1 - 2017
N2 - End-users compose ad-hoc business processes by integrating web applications to conduct online tasks. Generally, end-users have to enter information into web forms of web applications, and often repetitively type the same information into applications. It could be a tedious job for end-users to fill in web forms with identical information. To save end-users from repetitive typing and increase composition productivity, it is critical to propagate and pre-fill user inputs to web applications. In this paper, we propose an intelligent auto-filling framework collecting and propagating user inputs across web applications, identifying user usage patterns and contexts. The empirical results show that our framework, on average, achieves a precision of 74.5% and a recall of 58% on pre-filling web forms, and a precision of 82.25% and a recall of 68.4% on suggesting values to end-users if the end-users edit the initial pre-filled values.
AB - End-users compose ad-hoc business processes by integrating web applications to conduct online tasks. Generally, end-users have to enter information into web forms of web applications, and often repetitively type the same information into applications. It could be a tedious job for end-users to fill in web forms with identical information. To save end-users from repetitive typing and increase composition productivity, it is critical to propagate and pre-fill user inputs to web applications. In this paper, we propose an intelligent auto-filling framework collecting and propagating user inputs across web applications, identifying user usage patterns and contexts. The empirical results show that our framework, on average, achieves a precision of 74.5% and a recall of 58% on pre-filling web forms, and a precision of 82.25% and a recall of 68.4% on suggesting values to end-users if the end-users edit the initial pre-filled values.
KW - Business process integration
KW - Clustering
KW - Context-aware
KW - Usage patterns
KW - Web form auto-filling
UR - http://www.scopus.com/inward/record.url?scp=85015152569&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85015152569&partnerID=8YFLogxK
U2 - 10.1504/IJBPIM.2017.082747
DO - 10.1504/IJBPIM.2017.082747
M3 - Article
AN - SCOPUS:85015152569
SN - 1741-8763
VL - 8
SP - 16
EP - 30
JO - International Journal of Business Process Integration and Management
JF - International Journal of Business Process Integration and Management
IS - 1
ER -