IEEE Technology and Society Magazine - March 2021 - 93

Stage II: Processing the corpus of impact
measurement reports and populating the
ontology
The heterogeneity of the existing impact measurement reports will be the biggest stumbling block
to be overcome in this second phase. Also, the text
mining process will clearly be conditioned by the
results of the first phase of the research.
The objective of this phase will be to extract from
the measurement reports the necessary elements to
populate the ontology. Depending on the level of
detail chosen and the difficulty of extracting indicators from the corpus of reports, this populating process may be automatic or semiautomatic.
If possible, the system will extract all elements of
the report set without human intervention of any kind
and instantiate with them in the ontology. However,
a superficial study of the reports allows us to observe
the lack of easily identifiable patterns in them, which
will certainly make it difficult to fully automate this
task. In the event that it is not possible-in the scope
of this project-an automatic extraction, a semiautomatic settlement will be chosen, based on suggestions from the system but with a subsequent review
and final selection of human indicators.

Stage III: Development of a prototype CBR
system to support the measurement of the
impact of social enterprises
In the third phase of our research, we will build a
prototype of a CBR system focused on an area within
social entrepreneurship to be determined. Although
the objective of SIAMES in the medium term is to
serve as a support for any type of social project, at this
stage we have decided to focus our efforts on achieving a functional prototype that only covers one sector.
This prototype will have a project cases database
modeled according to the ontology. Using the similarity functions described in the next section, SIAMES
will be able to retrieve the cases most similar to that
of the end user by comparing the representation of
the new case with the existing case in the database.
As part of the research, we will also explore the possibilities of automatically generating this representation from a textual description of the project.
Social enterprises are under increased accountability pressures to demonstrate their effectiveness.
Emphasis is being made on measuring outcomes or
impacts and a growing portion of these organizations

March 2021

state that they need simple and standardized tools to
measure these effects.
Research on the assessment of social impact is
maturing. However, new challenges associated with
social enterprises arise, whose social impact assessment is still incipient. There is a need for consensus
on the definition and explanation of the interrelations between outcomes, indicators, and social
impact of a given population, a need to incorporate
standardized measures for concepts such as quality
of life, equity, social inclusion, and environmental
issues, among others.
This study deals with the development of a
methodology to evaluate the impact of social enterprises, based on the analysis of indicators related
to the social dimensions. The availability of subjective indicators of the changes to stakeholders as
a result of social enterprises activity are essential
to evaluate specific programs and their impact.
Only the availability of systematic key and relevant
data may facilitate making decisions that allow
for positive social change in the areas that affect
the population. Consequently, the main reason for
our study is to find a method for the selection of
indicators that allow to objectively verify the outcomes occurred for each stakeholders' group that
can be affected by a concrete social problem and
the changes that result of the activity of the social
enterprise.
To this end, in this research, we have proposed a
structure for ontology, pioneer in the field of social
entrepreneurship, we have presented an illustrative
prototype for the semiautomatic ontology population and, finally, we have exposed an original similarity function defined in the ontological space, the
core of the proposed CBR system.
With the combination of semantic text mining,
ontologies, and CBR systems, we have entered the
field of AI that combines symbolic and probabilistic
approaches, contributing with our results and experiences to the development of this current trend. We
have probed that this approach is valid for the resolution of the research challenge formulated, to apply
ontology-based semantic text mining methods on a
corpus impact measurement reports to subsequently
use the results in a recommender system applying CBR.
Furthermore, we have taken an important step by
applying AI techniques to the field of social entrepreneurship and impact measurement, an important

IEEE Technology and Society Magazine - March 2021

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - March 2021