Page 402 - Contributed Paper Session (CPS) - Volume 4

CPS2449 Louisa Nolan et al.
A web application was also developed for Optimus, which allows users to
explore the clusters and labels, and amend these if required. This enables the
data experts to carry out fast, intuitive quality assessment on the outputs, and
is an important component of assurance on this new dataset, constructed from
complex natural language processing algorithms. This is a good example of
how humans and artificial intelligence (AI) can complement each other.
Figure 4: summary of main words used by high growth companies
for different free text collected from their websites. Text in green is
more likely to be mentioned by high growth firms, whilst text in red is
less likely to be mentioned for the different free text entries.

Figure 5: dendrogram showing the result of the first iteration clustering
for a dataset of product descriptions using Optimus

4. Discussion and Conclusion
4.1 Discussion
In the faster indicators programme, we have supplemented official
statistics using novel data sources:

391 | I S I W S C 2 0 1 9