An Advanced ontology based automatic approach in improving the similarity by means of combining the sub ‐ graphs between the information

In this research an advanced and automatic approach in finding the new products which are similar to the previous one that results in a rapid experience in production of design of manufacturing procedure of the new product has been proposed. The proposed work is based on advanced ontology based semantic model which computes similarity between the sub-graphs in an effective manner. It builds a hierarchical structure by means of new similarity index that forms by overlapping sub-graph of existing two product concepts. By means of stored data the similarity measurement is calculated by matching the similar characteristics with the new one that helps in discovering knowledge. The examined result with the real-time data shows minimum computation cost along with high processing speed in similarity according to the global environment. Thus proves the proposed scheme is far better than other existing similarity approaches.


Introduction
* Generally information is the base of work in any simple business or dealing with a huge manufacturing industry. The simple question arises on the importance of information retrieval in running a business of all kinds. And the answer is information retrieval determines the time consuming and computation costs which results in the business hierarchy in a better way. According to the current industrial market there are several major drawbacks in handling the concepts of storing and reusing the manufacturing knowledge which are available in the companies.
To handle these issues and design a prominent solution in achieving the information retrieval, the best practicing method is ontology. Now-a-days information was handled in web based scenario and the question is how it is going to be handled in better way. Ontology can be built by certain representation of concepts which belongs to the domain. These representations are based on the impact of time, action, physical objects and various other factors. According to which ontology shares structural similarities and enables the process in an effective manner. In the real time environment, when an organization building a new factor in their business aspect, they need to learn the whole terminology even in which most of the time they will not have adequate knowledge on manufacturing techniques. This can be fulfilled by developing a best approach which will make it possible to perform the similarity of existing product with the new product in an automatic manner. By means of this novel approach, huge time will be saved with the concept of reusing the information that is already available in the organization's knowledge management. In addition, maintaining of huge information in an organization without the knowledge of existing data will result in memory consumption and drawbacks pertaining to maintenance in the overall level.

Related works
Due to the growth of various manufacturing industries and technological growth, there is an emergence of concept deals with automatic approach in finding the information of a management in an organization. Various research works are carried in developing an automatic analysis of available information and reusing it in a simple manner. According to (Bruno, 2015a), it is not simple to extract the definite manufacturing knowledge in their organizations. By the analysis of work on (Lowe et al., 2004) the designer took 20% of her time in searching and analyzing the appropriate information available in an organization.
In paper (Papakostas et al., 2010) it is shown by means of personally stored information, 40 % of required information can be gathered. According to the earlier employees worked on the enterprises, they had wide knowledge on documentation that will be helpful in future. The study on (Huang and Diao, 2008) proves that those documentations were difficult to understand and usage was not very prominent for the employees who will work in future, due to the documentation format as wells as their standards. In the papers (Denkena et al., 2007), due to the change business standards, about 50% of information was not presented in information system. Moreover the concept of information retrieval and reuse are not applicable for the systems, because they are only documented but not managed.
The evolving of semantic technology and design of knowledge management systems are helpful in realizing the details of companies in a minimum time consumption and reusing it effectively were shown in (Fortineau et al., 2012;Bruno et al., 2014). On this way ontology enables the possibility of combining information based on its abstraction levels and method of improvement in capturing as well as reusing (Chang et al., 2006). According to SMEs (Efthymiou et al., 2015), the knowledge management lack in performing the classification of past projects which take more time in identifying the similarity between new and past projects available in the memory.
From the work on (Liao et al., 1998;Bouzid et al., 2013) they show the similarities only on numerical attributes, otherwise on string attributes it based on edit distance functions. These works do not consider the meaning of the word and that is considered as a major drawback on these schemes. The Gene Ontology context (Teng et al., 2013) deals with similarity on bioinformatics which process the similarity between the genes. But it's not an effective model in achieving the efficiency in terms of medical diagnosis. On the paper (Groover, 2007;Bruno, 2015b), Groover discussed about the Fundamentals of modern manufacturing: materials processes, and systems in forming ontology for manufacturing process.
Several frameworks were developed by ontology using Jena, an open source Java framework for semantic applications. These tools were developed using APIs for RDF with SPARQL queries and supports Protégé ontology editor. In croft research work, the author demonstrates the semantic relationship by means of semantic network (Agosti and Crestani, 1995;Rau, 1987). According to Simone Ponzetto and Michael Strube they created a graph by representing inheritance between the words in Word Net (Knappe et al., 2003;Lopez et al., 2010). But these works require more time and possibility of mismatching, because there is no adequate information in analyzing the meaning of the word used as query.

Problem definition
In this section, we discussed about the various problem of manufacturing enterprises facing in the industrial market. In an organization the products were described by its product model which contains factors like technical data in order to produce the product in an appropriate manner. In detail the process model defines various data and information which gives important factors like processing time and requirements of the products. In other aspect the operations like dispensation, assembling, storing, managing, transporting and inspecting. The issues are on the basis of information on the management system and it may vary according to the product.
On retrieving the information, there is to be more similarity of new products with the existing one that will be more efficient than others. Here finding the similarity between the existing and new one took huge time consumption and computation cost. The important drawback is the existing system fails to consider the meaning of the word. It may lead to the possibility of finding similarity between the unrelated products. In addition, an information management fails to show the similarity between the products which may lead to load more information and that is already available in the management system. This increase data duplication and occupies more amount of memory space that also increase the time during the information retrieval process.

Proposed system
In order to overcome the issues, we involved in designing a supporting tool which indentifies automatically best product that is similar to the existing one which maximizes the speed of manufacturing new products. The mechanism of the proposed system is, by means of similarity index existing product is overlapped with new proposed which has maximum similarity. The ontology structure is formed on these bases by understanding the meaning of the word. In these manner subgraphs were created to know the best knowledge in understanding the information management system of an organization. Before creating our proposed frame, it is important to understand the concept of the proposed system. Fig. 1 shows the overall mechanism of the proposed system. As in Fig. 1, the ontology system separates system into function knowledge and manufacturing knowledge. The functional knowledge gives details about the product functionalities which is helpful to collect the information related to manufacturing.
By means of that, ontology creates mapping structure between the data and semantics tree structure. Based on the indentified functions, a link between the sustainable manufacturing and functions of manufacturing processes is created. The inte useful det manufacturi proposing indentified product wil available in Before d know about in an effect semantic m interoperab solution fo dealing wit Traditionally string-based Here it's for among the e is played by will form ma Fig. 2 sh such as Pro is not sema ontology by reference, b building the ontology dev Step 1: B the importa similarity is explained in (SKOS) (Kol Step 2

Framewo mputation
This section oposed fram mantic simil oduct. In nderstanding tology proce ocessing ope ocessing ope orking of pro b-graph were anner the ass sembling the anufacturing eated based b-graphs can ig. 4).   Fig. 4 shows the mechanism of proposed framework in finding the similarity between the two products. By which ontology creates tree and subtree matching the products similarity according to the given query. The new product to be manufactured is considered as product X. Next, the product undergoes process of identification automatically with class identification and match with the past product available in the information system. The semantic similarity is achieved with the product X by means of the achieved maximum similarity. The product X completes the manufacturing after several factors like grinding, shaping, cleaning, processing, annealing, surface processing and coating.

Sub-graph overlapping process
Fig. 5 shows the possibility of process which would be undergoing certain processes before completing its manufacturing stage. Various tree structures are formed with certain connectivity with one another. As discussed above, various factors such as grinding, shaping, cleaning, processing, annealing, surface processing and coating were assume with c1, c2, c3, …, cN respectively. Based on these relationships the tree structure is formed. Considering a new product to design, initially an organization has to check with the old information management. Based on the characteristic, the ontology creates a sub graph on the particular domain. From which the factors matching with new and existing data overlap each other in a prominent manner as shown in the figure. Thus according to the process, various sub-graphs such as G1, G2 and G3 are formed based on the product similarity achieved. Then by means of maximum similarity, union between the sub-graphs was formed automatically. The maximum similarity between the two subgraphs formed using the expression G1∩ G2.

Performance analysis
The semantic similarity is achieved by calculation based on the information content obtained and processed. By means of the Fig. 3 similarities are calculated using the expression (Eq. 1); ( 1, 2) = − ( , ) * (1) Here, D describes the similarity distance between the overlapped sub-graphs.
The D distance between two categories is calculated by the amount score gained between them. It can be shown as; • Identity: D(c1, c2)=0 ⇔ c1 =c2 • Normalization: 0 ≤ D(c1,c2) ≤ 1 • Triangle inequality: D(c1, c3) ≤ D(c1, c2) + D(c2, c3) • Next the information in the content analysis by using the expression (Eq. 2); where IC determines information content, p shows product and c describes the combination of factors involved for manufacturing the product as shown in Fig. 3. The proposed semantic similarity is achieved by (Eqs. 3-5); ( 1, 2) = ( , ) ( ) (3) The final result achieved is; ( 1, 2) = * ( ) ( 1, 2) = 1 ∩ 2 The above expression is computed based on product X which undergoes manufacturing. Before manufacturing the similarity is calculated with old information in the organization. By means of the expression the maximum similarity is achieved with the product X. In these various factors involved in manufacturing a product is considered, that is c1, c2, …., cN. Those are tabulated on the basis of information content (IC) best matching which is shown in Table 1 and 2. Here the calculation based on the sample product X undergoes various factors such as grinding, shaping, cleaning, processing, annealing, surface processing and coating and the values are tabulated. To analyze the proposed works efficiency the results are shown graphically in Fig. 6 based on the tabulation. In Table 2 it is clear that the new product X is matched with three existing ones and best one with maximum similarity is taken into the consideration.

Conclusion
In this paper we discussed about the problem of indentifying similarity between the existing and new product. By which a proposed model framework is designed using a modern approach of sub-graph overlapping based on maximum similarity. The information content and ontology formation between the products are achieved automatically, thus proving the effectiveness of our proposed work. The result and work shows that new product manufacturing is built with maximum speed in extracting the information, and forms product graph along with sub-graph proving as a prominent solution then the other existing works in the industry.