Zelenko, Dmitry, Oleg semin3
developed an automatic system that discovers companies which are in competition
from public information sources. In this the data is extracted and also uses
transformation learning techniques to get appropriate data normalization which
combines structured and unstructured sources uses probabilistic models to
represent the unlinked data and succeeds in discovering competitors. In this
authors also introduced iterative graph reconstruction process. The authors
also used machine learning algorithms for finding competitors. But this
technique has a problem of finding market demands.
Lappas theodoras, George Valkanas,
Dimitrios Gunopulos1 presented a formal definition of competitiveness between
two items. In this authors have used many domains and also handled the problems
in previous approaches. In this author consider the items are positioned in
multi-dimensional feature space and also considers the opinions and preferences
of users. However, the technique addressed many problems like finding top-k
competitors of a given items.
Pant , Gautam and Olivia RL Sheng
verifies that competing products are likely to have a similar web footprints a
phenomenon that refers to online isomorphism. In this they consider different
types of isomorphism between two firms such as overlap between the in-link and
out-link of respective websites. But the need for isomorphism feature limits
its applicability to products and makes it unsuitable for items and domains
where such features are not available (or) extremely sparse.
Mark and Margaret A.Petarat have suggested the frameworks for manually
identifying of competitors. Due to large and newly emerging of companies, it is
time consuming for us to find competitors manually.
Li, Rui, ShenghuaBao, jinWang, Yong Yo, Yn boacao accomplishes a task
for mining competitors with respect to an entity. Here entity refers to person,
product (or) a company. In this the authors developed an algorithm called
“CoMiner” which first extracts the comparative items of input entity and rank
them according to comparability. But CoMiner was developed for supporting a
specific domain and effort for further domains is still challenging.
Li, Rui, ShenguaBao ,JinWang, Yuanjie Liu, Yong Yu proposed a ranking
methods for finding competition information. In this they proposed a effective
techniques for finding competitors.