Technical reports
Interactive analytical modelling
May 2018, 142 pages
This technical report is based on a dissertation submitted December 2016 by the author for the degree of Doctor of Philosophy to the University of Cambridge, Emmanuel College.
Some figures in this document are best viewed in colour. If you received a black-and-white copy, please consult the online version if necessary.
DOI: 10.48456/tr-920
Abstract
This dissertation addresses the problem of designing tools for non-expert end-users performing analytical modelling, the activity of data analytics through machine learning. The research questions are, firstly, can tools be built for analytical modelling? Secondly, can these tools be made useful for non-experts?
Two case studies are made of building and applying machine learning models, firstly to analyse time series data, and secondly to analyse data in spreadsheets. Respectively, two prototypes are presented, Gatherminer and BrainCel. It is shown how it is possible to visualise general tasks (e.g., prediction, validation, explanation) in these contexts, illustrating how these prototypes embody the research hypotheses, and confirmed through experimental evaluation that the prototypes do indeed have the desirable properties of facilitating analytical modelling and being accessible to non-experts.
These prototypes and their evaluations exemplify three theoretical contributions: (a) visual analytics can be viewed as an instance of end-user programming, (b) interactive machine learning can be viewed as dialogue, and (c) analytical modelling can be viewed as constructivist learning. Four principles for the design of analytical modelling systems are derived: begin the abstraction gradient at zero, abstract complex processes through heuristic automation, build expertise through iteration on multiple representations, and support dialogue through metamodels.
Full text
PDF (9.7 MB)
BibTeX record
@TechReport{UCAM-CL-TR-920,
author = {Sarkar, Advait},
title = {{Interactive analytical modelling}},
year = 2018,
month = may,
url = {https://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-920.pdf},
institution = {University of Cambridge, Computer Laboratory},
doi = {10.48456/tr-920},
number = {UCAM-CL-TR-920}
}