BIPARTITE GRAPH FOR SEARCHING CHEMICAL

COMPOUNDS

Ashwini. M. Chalawadi.
Dept. of M.C.A
Visvesvaraya Technological University Belagavi,
Karnataka-590 018.
[email protected]

Abstract

The chemical industry is robust, developing Compounds are represented in data sets by pH
out in different directions supported by an value that ranges from number 0 to 14 or by
effective growing research establishment and by chemical name or represented using chemical
tried and true manufacturing and marketing structure. Searching massive compounds from
organizations. Along with these positives, large data sets within fraction of time require
however, has come one of the great challenge of new searching methods. Different searching
organizing the chemical compounds and finding techniques are applied to find text materials
them abruptly. This data is in unstructured either on web applications or on desktop
forms. Searching a millions of compounds applications or on mobile applications[1]. Many
quickly is a challenging task. Different conventional searching algorithms are
algorithms are used for searching data and are implemented such as Brute-force algorithm,
efficient in different prospective. Bipartite graph Boyer–Moore string search algorithm for pattern
gives efficient solution for data mining matching. Efficiently mining the large scale of
applications. This paper presents a chemical compounds becomes an significant
unconventional algorithm for searching problem for the future development of the
chemical compound through the use of bipartite chemical industry including pharmaceutical,
graph. Bipartite graph gives fully optimized agrochemical, biotechnological, fragrances, and
solution to search any kind of compound by general chemical companies. Graph theory
accepting pH constant as an input. The proposed concepts give different approach to solve many
approach is time efficient because of fewer complex applications.
number of comparisons required in searching. A graph (undirected graph) G consists of
non-empty set V of points (Vertices or nodes)
Keywords: Graph, Graph Traversal, bipartite and set E of unordered pair of elements in V
graph, complete bipartite graph, pH value, (called edges). The graph G is denoted by
Chemical Compound. ordered pair (V, E) i.e. G (V, E) or G: (V, E).
Graph traversal is visiting all the nodes in
a graph in particular manner. A bipartite
graph ( bigraph) is a graph whose vertices can be
I. Introduction divided into two disjoint sets and (that
The increasing number of chemical is, and are each independent sets) such
compounds in various chemical industries and that every edge connects a vertex in to one
life science industries requires the development
in . Vertex set and are often denoted
of new methods and techniques for their as partite sets.
handling. Chemical industries come across
many challenges of visualization of millions of
compounds. Over the period of time there
actually has been a incredible increase in the
number of available chemical activity, leads to
meticulous mining of compounds in a
standardized form . Most of the chemical

978-1-5386-7709-4/18/$31.00 2018
c IEEE 71
 II. Related Work

Fig 1: Complete Bipartite Graph

A complete bipartite graph is

a graph whose vertices can be partitioned Fig. 3: Graph G for chemical compound
into two subsets U and V such that no edge representation with respect to pH values.
has both endpoints in the same subset, and
pj ={ pj1, pj2,……pj14} is the set of pH
every possible edge that could connect
constants of chemical compounds,
vertices in different subsets is part of cj={cj1,cj2,…cjn} is the set of chemicals. An
the graph. i.e. every vertex of the first set is edge ej exists if chemical is found and pj=cj.
connected to every vertex of the second set. An edge ej signifies the association between
pH, a quantitative measure of the acidity or pH constant and chemical. By bipartite
basicity of aqueous or other liquid solutions. graph {pj1 ∩pj2∩pj3∩pj4∩pj5…..pj14}=ø. and
Acids and bases are two groups of chemical {cj1∩cj2∩j3…∩cjn}=ø. Means there is no
edge present between pj1 ,pj2..…pjn. and also
compounds that generally have inconsistent
between cj1, cj2 …cjn. i.e. by bipartite graph
characteristics. Chemical compound pH is there is no edge exist between same set of
value ranges from 0 to 14. A compound with vertices.
pH value zero is neutral compound. A
compound with the pH value less than 7 is So this in turn reduces complexity of
comparison. Here pj0=0 pj1=1, pj2=2….. pj14= 14
acidic and a compound with pH value
and
greater than 7 is basic compound. These cj0= {Chemical compound start with pH value ‘0’}
numbers describe the concentration of c j1= { Chemical compound start with pH value ‘1’}
hydrogen ions in the solution and increase
c j14= { Chemical compound start with pH value ‘14’},
on a negative logarithmic scale.
{c j1,c j2 ,c j3,…..cj14} are subsets of cj.

The following algorithm searches a

chemical compound, based on pH value of the
chemical constant.

ALGORITHM: Bipartite Graph Search

begin
Fig. 2: pH scale for chemicals
visit pj
compare pj==cj
The contributions of this work are as
if pj = cj
follows. Section II describes the proposed
add an ej
work, Section III presents possible outcome
traverse pj to cj1…cjn.
of the work and Section IV concludes the
return cj
paper.
else if
return -1
else if

72 2018 International Conference on Computational Techniques, Electronics and Mechanical Systems (CTEMS)
 repeat pj compounds in set cj . 14 is the total no. of pH
traverse pj to cj. constants. While during the second time
return cj traversing of the graph no comparisons found
else i.e. n=0 comparisons. Directly visits the vertex
return -1 set pj and returns the matched set of compounds.
So in second time traversing the graph requires
end approximately ‘t/2’secands. Time complexity of
algorithm is explained in the following Table 1.
Fig 3: Bipartite Graph Search Algorithm
Vertex of set pj connects to different Table 1: Time Complexity of algorithm.
vertex set of cj by pH values of the chemical Traversal Total Total no. of Time
compound with its respective compound by an No. chemical comparisons in
edge ej by comparison pj= cj. If pj = cj an edge compounds sec.
occurs, that represents an association between |cj|
pH constant and chemical compound and 1 n 14*n t
returns the set of related compounds cj., Suppose
2 n 0 t/2
if pj≠ cj an edge does not exist, that represents an
inassociation between pH constant and chemical
compound and returns -1. The bipartite graph structure can be easily
The above process of traversing the adjusted with the addition of new vertices
vertex set of pj and vertex set of cj by input as a (chemical compounds) and edges. Moreover
pH constant belongs to set pj by comparison of this algorithm can be parallelizable and work
pj=cj and forming edge, happens in first time in a dynamic context of stream of
traversing of the graph G. documents.
The proposed algorithm is different
III. Result
from existing algorithms like Brute force,
During the first graph traversal the
comparison takes place between pj and Boyer–Moore search algorithm and other
{cj1,cj2..cjn} € cj . An edge occurs between pH search algorithms. Those are used for single
constant and chemical compound belongs to set word search and multiword (pattern
cj. An edge ej exist only if pj=c j. Suppose if pj≠cj matching) search. Therefore, time
an edge does not exist, that represents an in complexities of existing algorithm have not
association between pH constant and chemical been compared with proposed algorithm in
constant .However if chemical compound does above table 1.
not exist in the set cj returns -1.When an pH
constant is given as the input that belongs to set IV. Conclusion
pj, in first traversal (operation). During the The proposed bipartite graph algorithm
second graph traversal, only the traversing takes for searching a chemical compound is modern
place between pH constant and respective approach for minimum time complexity. The
chemical compound, through edge , no graph representation is a simple structure that
comparisons takes place because in first graph can be integrated in dynamic methods. Due to
traversing only edge is formed between pH simple structure new vertices (chemical
constant and associated chemical compound constants, chemical compounds), edges and sub
and returns the respective compound. So pj is graphs can be added with less or no complexity
entered second time (i.e. pH constant is of adding data . The proposed algorithm uses
repeated), only traversing takes place between undirected graph so there is no restriction to
pH constant and chemical compound no edge traverse the graph between to one vertex to
formation occurs. i.e. no comparison between pj another vertex. The algorithm works efficiently
and {cj1, cj2, cj3,….cjn}. So this minimizes the on ordered as well as unordered data.
comparisons. If there are n compounds, during
the first time traversing of the graph it takes
14*n comparisons and time required is ’t’ sec.
‘n’ represents the total no. of chemical

2018 International Conference on Computational Techniques, Electronics and Mechanical Systems (CTEMS) 73
References
[1] Thiago de Paulo Faleiros and Alneu de
Andrade Lopes, “Bipartite Graph for Topic
Extraction”, Proceedings of the Twenty-
Fourth International Joint Conference on
Artificial Intelligence (IJCAI 2015),
pp.4361,4362, 2015
[2] Lihui Lan et al. “Anonymizing Social
Network using Bipartite Graph”, 2010
International Conference on Computational and
Information Sciences, pp. 994, 2010.
[3] Yousef Abuzir and Thabit Sabbah
Journal of Computer Science
Volume 8, Issue 1,pp.61-67,2012.
[4] LiminFeng and Yan Yang ,” Bipartite
Graph-based Keyword Query Results
Recommendation”, 2013 International
Conference on Mechatronic Sciences, Electric
Engineering and Computer ,Dec 20-22 ,
Shenyang, China (MEC),pp.1584-1585, 2013.
[5] Yousef Abuzir and Thabit Sabbah ,Title not
found, Journal of Computer Science
Volume 8, Issue 1,pp, 61-67,2012.
[6] Ankit R. Chadha et al.” Modified Binary
Search Algorithm”, International Journal of
Applied Information Systems”, Volume 7– No.
2, pp. 37-38, April 2014.
[7] Pengfei Xu et al. “Detecting Objects in
Image Collections Using Bipartite Graph
Matching” , 2012 International Conference on
Systems and Informatics (ICSAI 2012),pp.
1906, 2012.
[8] R.Nigel Horspool el al.”Constructing Word-
Based Text Compression Algorithms”, pp. 1-2.

74 2018 International Conference on Computational Techniques, Electronics and Mechanical Systems (CTEMS)