Enterprise Relationship Network: Build Foundation for Social Business

Liqiang Wang, Shijun Liu*, Li Pan, Lei Wu, Xiangxu Meng

School of Computer Science and TechnologyShandong University

Jinan, 250101, P.R.Chinawanglq1989@163.com, {lsj, panli, i_lily, mxx}@sdu.edu.cn

Abstract—��Social business moves beyond linear, process-drivenorganizations to create new, dynamic, networked businessesthat focus on customer value. Enterprise relationship network(ERN) can be used to support social business by maximizingcurrent and future opportunities and facilitate network-enabled processes, which can lead to value co-creation. In thispaper we give the specification of ERN, which links the mainentities in social business together, such as enterprises,business activities, employees and products. ERN provides aset of methods for analyzing the structure of whole entities aswell as a group of algorithms for exploring the patterns inthese structures. We present the technique architecture ofERN and describe how the ERN supports social business. Wecan get a lot of valuable information from ERN, which can beused in enterprise management, employee collaborations andnetworked businesses. At last, through a case study on theplatform of SDCMSP, we evaluate how our proposed approachsupports social business and show some relationshipvisualization results.

Keywords- enterprise relationship network; socail network;social business

I. INTRODUCTIONSocial business refers to that organizations consciously

use social tools, social media and social networking,prepensely integrate Web2.0 technology and Internet spaceto rebuild the communication channel to consumers andchange the way of organization management and businessoperation. The next generation of market-leadingorganizations will digitize their enterprise model with newcapabilities enabled by social technologies. Social businessanswers to the major technological and economic shifts thatdefine innovation today. Social ways of working harness theexplosive growth of mobile, cloud and big data, and serve asthe foundation for effective employee and customerengagement. As of September 2013, there are 14,693,100enterprises in China which is really a huge space for socialbusiness. The key to realize social business is thecomprehensive utilization of social information whichapparently is in large quantity and diversified. The role ofinstantaneity in social business is very critical for theenterprise to maximize its value to the business. Hence, bigdata is fitful to support social business.

* To whom correspondence should be addressed.

Most enterprises are not willing to use public onlinesocial networks (OSNs) for the considering of privacy andsecurity. Therefore, enterprise social network (ESN)emerged and grew quickly, some software based on ESN hascome out, e.g., Salesforce’s Chatter, Microsoft’s Yammerand IBM Connections. This software provides conveniencefor collaboration and communication among employees andalso between enterprises in premise condition of theguarantee of privacy and security. But in these applicationsESN is confined to communication among people ignoringthe business which is the core factor of enterpriserelationships. Hence, in [1] we proposed a new ERN model,in which we built the network of enterprise relationshipsregarding the enterprise as the core entity and the businessbetween enterprises as the principal line. The ERN model isexpressed by weighted directed graphs and hypergraphs, itconsists of four graphs, the enterprises graph (ENTG), theemployees graph (EMPG), the products graph (PROG) andthe three-level hypergraph (TLHG). The comprehensiveutilization of these graphs can link the main entities in thebusiness together, such as enterprises, business, employees,products etc. Our proposed ERN model has broadapplication prospects. For the similarity with social network,many algorithms and methods in social network analysis(SNA) can be applied in ERN.Enterprise Relationship Management (ERM) is basically

a business strategy based on the leverage of network-enabledprocesses and activities to transform the relationshipsbetween the organization and all its internal and externalconstituencies in order to maximize current and futureopportunities [2]. In social business, ERM is a critical entrypoint to tease the relationships clearly. We hold the opinionthat ERM is made up of Supplier Relationship Management(SRM), Customer Relationship Management (CRM),Product Relationship Management (PRM) and HumanRelationship Management (HRM).The remainder of this paper is organized as follows. In

sectionⅡ, we introduce the related work which is the resultsof other people mainly about the theory and analysis ofsocial network and Big Data. We refine the definitions andmodel of ERN, more details are provided in section Ⅲ. OurERN model is applied to support social business which isdiscussed in section Ⅳ. In section Ⅴ, we present ourexperiment and relationship visualization based on theexperiment data. Finally the conclusion and future work aregiven in sectionⅥ.

2014 IEEE International Congress on Big Data

978-1-4799-5057-7/14 $31.00 © 2014 IEEE

DOI 10.1109/BigData.Congress.2014.57

352

2014 IEEE International Congress on Big Data

978-1-4799-5057-7/14 $31.00 © 2014 IEEE

DOI 10.1109/BigData.Congress.2014.57

347

2014 IEEE International Congress on Big Data

978-1-4799-5057-7/14 $31.00 © 2014 IEEE

DOI 10.1109/BigData.Congress.2014.57

347

2014 IEEE International Congress on Big Data

978-1-4799-5057-7/14 $31.00 © 2014 IEEE

DOI 10.1109/BigData.Congress.2014.57

347

II. RELATED WORKThere are some distinctions of ERN compared to existing

social networks and enterprise web applications, such associal network, Enterprise 2.0 [3] and enterprise socialnetwork [4]. The core factor of social network is people, andthe social relationships are usually not related to business.Enterprise 2.0 is the use of emergent social softwareplatforms within companies, or between companies and theirpartners or customers. It aims to help employees, customersand suppliers collaborate, share and organize information viaWeb 2.0 technologies [5]. The typical application ofEnterprise 2.0 is web enterprise software such as CRM andSupply Chain Management (SCM), which focus on business,but compared to ERN, this software is most limited in one orsome enterprises. Enterprise social network is still thenetwork of people, in which employees may collaborate on abusiness by sharing some information, but the informationmay be inaccurate even not veracious.Some excellent theories in social network modeling are

very useful. Holroyd examined the general characteristics ofnetworks and the factors which influence their stability andgives particular attention to the idea of 'coupling strength'and 'connectivity' between elements of a network, whichprovided the experience for our ERN model [6]. Junhua, D.et al. extended a high level Petri nets with channels forformally modeling social networks [7]. And Conti, et al.presented a model of the ego network by correlating the levelof emotional closeness of a social relationship to the timeinvested in it [8].There are some methods to build an employee social

network, Jin, et al. combined the organization graph and thesocial interaction graph to analyze and model user interactionin enterprise social networks [9]. DiMicco, J.M. et al.designed a social network site to support employees withinan enterprise in communicating with each other throughpersonal and professional sharing [10]. Bennett gave thebenefits of communicating and collaborating in a real-timeenterprise social network [11]. Kohout & Neruda introducedan algorithm for social network graphs clustering, which canbe applied in ERN [12]. Dong, L. et al. uses an expertfinding model to carry on the influence analysis [13].SNA has been a hot topic and got a number of

achievements, if these achievements were applied into ERN,we would get many interesting and valuable results. Jamali,M. et al. presented a state of the art survey of the works doneon social network analysis ranging from pure mathematicalanalyses in graphs to analyzing the social networks insemantic Web [14].Big Data provides a powerful tool to support social

business. Singh, S. and Singh, N. presented that Big Dataprovides an opportunity to create unprecedented businessadvantage and better service delivery, which also requiresnew infrastructure and a new way of thinking about the waybusiness and IT industry works [15]. Because of the differentcharacteristics of Big Data with traditional data, we need anew way to analyze and visualize the data. Jinson, Z. & MaoLin, H. presented a 5Ws model to process Big Data analysisand visualization by attributing the large and various data

with six dimensions, and they did some works on clusteringvisualization with this model [16].

III. ENTERPRISE RELATIONSHIP NETWORK MODEL

A. The Entities Of ERNWe presented a hypergraph-based hierarchical ERN

model in [1]. The ERN is like most of the social networks.But something is different in ERN. There is only one kind ofentity in most social network, which is always the person. InERN, there are multiple kinds of entities, such as enterprises,products and employees to model the ERN and connect allnecessary factors in business. The following is the factorsappeared in ERN.Enterprise (ENT): Enterprise is the core entity in ERN,

which is the base of relationships. Here enterprise may be aparent company or a subsidiary which is the smallest unitwith the ability to build relationships to other enterprises. Touniquely identify an enterprise, we choose the ChinaOrganization Code as the ID of an enterprise. As anotherway, Data Universal Numbering System is widely usedinternationally.Employee (EMP): Every business and routine is finally

performed by employees. Although enterprise is the coreentity of ERN, employee plays an important role inconnecting the business process. Every business in ERN is incharge of an employee. In addition, relationships betweenemployees can be built for communicating with each otherconveniently. These relationships finally form the employeesocial network, which has been extensively studied.Product (PRO): products are needed to be uniquely

identified in ERN. Hence, we should manage the informationof the products in a clear way. In ERN, there are severalproducts in every enterprise. As a material supplier, theenterprise proposes its products to ERN, and publishes thisinformation to other enterprises, which plays a role ofadvertisement.Purchased items (PUR): The enterprise producing and

selling products needs to purchase materials and parts fromother enterprises. The enterprise should propose the bill ofmaterial (BOM) for every product. Through this information,we can get the procurement demand of an enterprise. Asupplier can link its products to the purchased items ofanother enterprise to show the desire for establishingbusiness.

B. The Relationships Of ERNWe refine the ERN model in a more formal and detailed

way. Figure 1 [1] is a three-level hypergraph. TLHG = (ENT,EMP, PRO, H). TLHG is a directed hypergraph. ENT is theenterprise node collection, EMP is the employee nodecollection, and PRO is the product node collection. H is thehyperedge collection. � Hi∈H, Hi = (ENTi, EMP’, PRO’, p),ENTi∈ENT, EMP’ � EMP, PRO’ � PRO, and p is thevector of parameters for the relationship. A hyperedge inTLHG means that enterprise ENTi has employees in EMP’and products in PRO’. The TLHG is a comprehensivedirected hypergraph which links ENTG, EMPG and PROGtogether.

353348348348

Figure 1. Hypergraph-based hierarchical ERN model

The middle layer is enterprise graph. ENTG = (ENT, E).ENTG is a directed enterprise graph. ENT is the nodecollection of the graph. � ENTi∈ENT, ENTi is an enterprise.E is the edge collection. � Ei∈E, Ei = (x, y, p), x, y∈ENT,and p is the vector of parameters for the relationship, e.g., theparameter k and s of p respectively determines the kind andstate of the relationship. An edge in ENTG means that abusiness relationship exists between enterprise x and y.The bottom layer is employee graph. EMPG = (EMP, E).

EMPG is an undirected employee graph. EMP is the nodecollection of the graph. � EMPi ∈ EMP, EMPi is anemployee. E is the edge collection. � Ei∈E, Ei = (x, y, p), x,y ∈ EMP, and p is the vector of parameters for therelationship, e.g., a parameter of p is k which determineswhether the edge is within an enterprise or not. And anotherparameter s determines the state of the relationship. An edgein EMPG means that employee x and y have a socialrelationship. We use solid edges to link the employees withinan enterprise and dotted edges to link the employees betweenenterprises in EMPG.The top layer is product graph. PROG = (PRO, PUR, F,

E). PROG is a combination of BOM trees and a directedgraph. PRO and PUR are the node collection. � PROi∈PRO,PROi is a product. � PURi∈PUR, PURi is a purchased item.F is a forest of BOM trees. � Ti∈F, Ti = (PROi, PUR’, TE),PROi∈PRO, is the root node of Ti. PUR’ � PUR. TE is thetree edge collection, � TEi∈TE, TEi = (x, y, p), x∈PUR’∪{PROi}, y∈PUR’, p is the vector of parameters for the treeedge. An edge in Ti means that the product PROi has a

purchased item in PUR’ or a purchased item has a smallerpurchased item. E is the edge collection of the directed graph.� Ei∈E, Ei = (x, y, p), x∈PUR, y∈PRO, and p is thevector of parameters for the relationship. A directed edge inPROG means that the source of the purchased item x isproduct y.BUS is a group of edges of the four graphs with

parameters. BUS = (ENTG, EMPG, PROG, TLHG, B). B =E(ENTG)∪E(EMPG)∪E(PROG)∪E(TLHG). � BUSi∈B,BUSi = (ENTG’, EMPG’, PROG’, TLHG’), ENTG’ �E(ENTG), EMPG’ � E(EMPG), PROG’ � E(PROG),TLHG’ � E(TLHG). Here business is not completely thesame as the real business between enterprises. Though BUSimay be an incomplete business process, it can link all thefactors in part of the business process, through which wecould describe the business clearly. BUS helps us tounderstand the business, e.g., we can get which employeesparticipate in the business and what enterprises and productsare involved in and what are the relationships among them.

C. The Characteristics of ERNERN makes enterprise as the core entity and the business

between enterprises as the principal line which is differentfrom social network. The core entity is enterprise in ERNinstead of people in social network, and the relationships arebuilt according to business but not only whether there iscommunication between the entities or not. ERN integratesenterprise, business, product and employee that the basic andnecessary elements in the process of carrying on the business.And the enterprise business relationships are the principalline which runs through the entire business process. ERN issimilar to social network in some aspects, consequently mostmethods of SNA are available in ERN.The complexity of our ERN model is mainly the

complexity of algorithms in graphs and hypergraphs. Tobuild a graph, the complexity is O (V+E), V is the number ofthe nodes and E is the number of edges, so the complexity isO (n2), n is the number of the nodes. The complexity of thethree-level hypergraph is O (n*m*k), n is the number ofenterprises, m is the number of employees and k is thenumber of products, so that the complexity is O (n3). Hence,the complexity of our ERN model is O (n3).

IV. SUPPORTING SOCIAL BUSINESS BY ENTERPRISERELATIONSHIP NETWORK

In today's competition, grasping the information is thekey to success. But the information explosion makes it hardto utilize the information comprehensively. An enterprisewould gain powerful advantages over the market competitorsif it could leverage all the information with the wholeavailable data rather than just a part of it. Big Data can helpto get insights and make better decisions.Social business is a promising and effective way beyond

all doubt for its comprehensive utilization of information.To support social business, our ERN model integratesvarious data and organizes it in a formal way for moreconvenient analysis. And we use technologies of Big Data to

354349349349

store, transmit and analyze the social media produced inERN and external.

A. Technology ArchitectureFigure 2 illustrates the technology architecture of ERN to

support social business, which regards the ERN as the coreinformation organizing layer.

Figure 2. Technology architecture of ERN.

1) Infrastructure layerThe bottom layer is the foundation to realize intelligent

deployment and data storage and transmit to support socialbusiness. The hybrid cloud platform provides the basicsupport to realize intelligent deployment, distributed dataengine and data warehouse. Part of the hybrid cloud is ourprivate cloud platform, and another part is some publicclouds to support applications based on public cloud. It’snecessary for an enterprise to make various existing anddeveloping systems work together and collect social mediato support social business. The ERN platform includes manyfunction modules and will support many differentapplications to realize social business, which needs moreintelligent deployment to deploy and manage theseapplications. The intelligent deployment should consider notonly the performance requirement of the applications butalso the security. The characteristic of social media isvolume and variety so that how to store the data is a problem.The data warehouse belongs to schema-free design whichhas the advantage in enabling applications to quicklyupgrade the structure of data without table rewrites and alsoallows for greater flexibility in storing heterogeneouslystructured data [17]. The data source of the Big Data isusually extremely widespread so that it needs to extract and

integrate the heterogeneous data. The distributed data engineis responsible for efficient data processing to meet the speedrequirement. The data engine provides efficient data accessApplication Programming Interface (API) to applications.2) Enterprise relationship network layerThe core of our architecture is the Enterprise relationship

network layer. It includes several important components.Social networking. To realize social business, the

enterprises, the employees in one or some differententerprises and the customers should communicate with eachother adequately on the premise of the guarantee of privacyand security, which will strengthen partner and customerrelationships. The social networking includes not onlyenterprise social networks but also public OSNs with somerestrictions, which are not directly used by enterprises butprovide the information for data analysis.Social analytics. Social analytics provides the ability to

mine the valuable information out from the volume andseemingly useless data.Social collaboration. ERN provides the convenience for

employee and business collaboration. The employeecollaboration is based on employee social network in ERN.It improves the office efficiency greatly that employees caninteract with each other by sharing information and findexpertise knowledge through the employee social network.And business collaboration is based on the whole ERN,which needs to consider all the factors in business in order tocreate stronger operations.Social content. Social content includes various

structured and unstructured data, e.g., sensors, documents,audio, video, pictures, email, business data, socialrelationships information, etc. The social contentmanagement is based on data warehouse and distributed dataengine to store and process the data more efficiently.Visualization. Data visualization is a good tool for

people to understand and analyze the vast complicated data.According to the characteristics of social content, we cantake some special ways to visualize the data, through whichpeople will get the valuable information from the enterpriserelationships intuitively.3) Applications layerThe top layer is applications based on ERN to support

social business. The goal of this layer is to help peopleintegrate the information together, carry on the data analysisand support business and people collaboration. We cancreate a smarter workforce by providing an integratedsmarter workspace for employees to fully accelerate the flowof information and knowledge and eliminate informationbarriers.

B. Social Analytics In Socail BusinessBig Data analytics is a key enabler for social business

due to its advantage in processing volume and various datarapidly. We can use Big Data technologies to carry on thesocial analytics, which helps business intelligentlyunderstand social media and promptly respond to emergingopportunities. Because of the new characteristics andrequirements, some traditional analytics techniques like data

355350350350

mining, machine learning and statistical analysis shouldadjust to the big-data era.Firstly, through the links in ERN, we can get the

employee, enterprise and product information and therelationships among them in a business process. With thehelp of new visualization technologies, we can show theserelationships more precisely. This information helps torealize the business collaboration within an enterprise orbetween enterprises.Secondly, we can expand valuable relationship in a wider

enterprise network. For example, enterprises usually knowtheir direct suppliers only. But with the help of ERN, we canrealize optimized multi-echelon supply chain deployment.Thirdly, some useful implied relationship can be mined

by the social analytics. Sociologists identify strong and weakties as the main types of relationships in a social network.This phenomenon also appears in ERN. For example, byanalyzing the supply relationship and product relationship,an enterprise can find its potential competitors or partners.

C. Applications Collaboration Based On ERNERN has broad application prospects for supporting

social business. Based on ERN, we could develop theapplication of the ERM. SRM manages the supplierrelationships of an enterprise which is similar to SCM, butSRM puts the suppliers in the whole relationship networkand pays more attention on the combination of relationshipsand business. CRM based on ERN can take good use of theinformation of people, products, enterprises and relationshipsamong them, hence it is easier to manage the customerrelationships and find the new opportunities. PRM plays anassist role in ERM to help understand the relationships ofproducts which link the enterprises together. It’s easier for usto discover more valuable information after mastering therelationships of products. HRM is based on employee socialnetwork in ERN. Employee communication andcollaboration can be realized based on HRM, which willimprove the office efficiency and reduce the overhead cost.Our ERN platform provides a set of API to support the

applications based on ERN. The API takes the technology ofWeb Service which can be called by the applications flexiblyand conveniently. The applications can use the original datain ERN to finish some business process or carry on somebasic analysis. And they can also directly get the resultsanalyzed by platform through Big Data technologies. Theapplications will be various and excellent with the ampledata source from ERN such as enterprise relationships,product relationships, employee relationships and all kinds ofsocial business media collected or produced by ERN. Animportant part of ERN is employee social network, based onwhich we can develop applications to supportcommunication and collaboration among colleagues in oneenterprise or employees in other enterprises with restrictionsto guarantee security and privacy. Such applications likeHRM will be popular for its openness andcomprehensiveness.

V. CASE STUDYIn this section, we introduce a case study based on

Shandong Cloud Manufacturing Service Platform(SDCMSP). SDCMSP concentrates on manufacture inShandong province, China, but not limited in this industryand geographic area. The core idea of this platform is ERN,which all of the functions of SDCMSP relies on.

A. Generating DatasetSDCMSP has been implemented, but not yet formally

put into use. Consequently we made a dataset trying tosimulate the real situation. In this simulated dataset, we justconsider one kind of relationship, the supply-purchaserelationship.We add five thousand supplier enterprises and fifty

purchaser enterprises, based on which we build aboutseventy thousand business relationships. These relationshipsare made according to some certain rules. The fiftypurchasers are the core enterprises, to which otherenterprises provide materials and products. Each of the corepurchasers has 1 to 3 products, and every supplier has tenproducts, so totally there are about fifty thousand products.For a core enterprise, each product has 300 to 500

purchased items, and we randomly select 1 to 3 suppliersfrom the five thousand suppliers for each purchased item. Onthe whole, every core enterprise has 600 to 1,800 suppliers.We select 500 core suppliers, apply one purchased item foreach product and randomly select another supplier to supplyits product for the purchased item, and finally we get aboutfive thousand relationships.

B. Relationship Visualization By A Directed GraphThe data and relationships in ERN seem to be fogged

when the scale reaches a certain level. Hence, relationshipvisualization becomes quite necessary for people to deeplyunderstand and analyze the data in ERN. Firstly we use thetraditional directed graph to show the relationships.Figure 3 is a directed graph based on the supply-purchase

relationships. This picture is the result after fading treatment,before which we can hardly see anything but a black areadue to the large amount of nodes and edges. Through thispicture we can get the information of all the enterpriserelationships in ERN, by which we can do some simpleanalysis, e.g., we can find out the enterprise with the mostsuppliers and the supplier with the most purchasers. Figure 3is an overall picture so that we can hardly see the details.Then we can zoom in, and the core enterprises are extremelyobvious in figure 4. In this picture, we just consider supply-purchase relationship, and we need a new way to visualizethe relationship when more kinds of relationships are joinedinto the dataset, e.g., we can use different colors to representdifferent kinds of relationships and apply a layered structureto separate these relationships. Traditional visualization waysare not enough to show the massive and various relationshipsin ERN.

356351351351

Figure 3. The white dots are enterprises and the directed edges arerelationships from suppliers to purchasers.

Figure 4. When we zoom in to figure 3, we can pick out core enterprisesmore clearly.

C. Chord Diagram With Interactions To Show TheRelationshipsPreciselyFigure 3, 4 are traditional ways to visualize the enterprise

relationships. This way is more fitful to show the overallfeatures. But when trying to find more details in this way, wecan realize that it’s a little hard to analyze the relationshipswith the chaotic nodes and massive directed edges. So weintroduce another way to show the relationship informationmore clearly. We can see the chord diagram in figure 5, thenodes are located at the outer circle and the edges at the inner.And we add interactions that when the mouse moves over anenterprise all the related relationships are highlighted. Thein-edges are red and the out-edges are green, which canavoid the confusing arrows. And we also bind the edges in

some degree. By the chord diagram we can see the exactrelationships of every enterprise and get the regularities ofrelationship distribution. One disadvantage of chord diagramis that the outer enterprise should not be massive. Thispicture consists of one hundred enterprises, and if weincrease the cycle diameter and reduce the word size, thenodes number could increase to about seven hundred. Wecan use enterprise classification to cover this shortage. Everytime we just show the enterprises that we are most interestedin. In this picture we only show some core suppliers. And wecan see that every core supplier nearly has the same amountof relationships. The reason is that we randomly built therelationships when generating the dataset.

Figure 5. Chord diagram with interaction. The outer cycle texts areenterprise nodes and the blue edges are enterprise relationships. Thehighlighted edges are the relationships of one selected enterprise.

D. Supporting Multi-echelon Supply Chain DeploymentFigure 6 shows the supply chain of an enterprise. The red

edges are the supply relationships of a supplier of the coreenterprise. We can get the information from this figure thatsome black enterprises supply to the green selectedenterprise, and this green enterprise supplies to the coreenterprise as well as the other green suppliers. Through thisnetwork we can deploy the two-stage supply chain, ifretrieve all the enterprise, we could get multi-echelon supplychain.Combined with PROG, the multi-echelon supply chain of

a kind of product is easily to get, which helps the coreenterprise to reduce the cost caused by multi agent throughpurchasing the product directly from the original supplier. Aswe get the multi-echelon supply chain, it’s more convenientfor the core enterprise to control the risk of stockout. OurERN platform provides the credit rating and operation statusof an enterprise, which can be easily got by analyzing thebusiness data using Big Data technology. The core enterprisecan analyze the supply risk of its suppliers with the

357352352352

information of multi-echelon supply chain and estimateinformation from the platform. If the suppliers of a first-tiersupplier operate not so well recently, the stockout risk willincrease. Then the core enterprise should take some actionsto cope with the risk, e.g., it can increase the ratio of thepurchase from other better suppliers.pp pp

Figure 6. Supply chain of an enterprise. The biggest red sphere is the coreenterprise. The green disks are the suppliers of the core enterprise. Theblack disks are the suppliers of the green enterprises and not of the coreenterprise. A directed edge is from the supplier to the purchaser.

E. Recommanding Potential Suppliers And CompetitorsRecommendation is always a commercial appreciation

service for its effect of advertisement. In ERN we can masterthe information of every enterprise, product and businessprocessed through ERN, which can be used to run someanalysis algorithms to recommend enterprises, products orservices to target customers [18].Figure 7 is a part of PROG, which helps to understand

the source of the parts of a product. PROG provides theinformation of the product relationships among enterprises.Combined with figure 3, we can recommend the productsand enterprises to the target enterprises. Through PROG wecan know whether a product and another are the same kind,and through ENTG we can realize that on this kind ofproduct which supplier is most popular and whose productsare of most high quality so that we can recommend thepotential supplier and its products to the target enterprises.As a supplier, it may consider more about potential

competitors to get more business opportunities and enhancecompetitiveness. And similarly with potential supplierrecommendation, we can recommend the potentialcompetitors to an enterprise. Through ERN we can knowthat which products are the same kind, which enterprisesproduce these products and which enterprises these supplierssupply these products for. The enterprises which supply thesame kind of product come into competition. And then theenterprise can search more information about competitors totake actions to enhance competitiveness. The recommend

algorithm should consider more factors, e.g., thegeographical position and credit rating of enterprises, therestriction of the parts of a product, the business history orsocial contact to the target enterprises etc.

Figure 7. Links between purchased items of a purchaser and products ofits suppliers. The biggest red sphere is the purchaser. The smaller red disksare the suppliers. The green disks are the products of the suppliers. The twogreen spheres are the products of the purchaser. The black disks are thepurchased items of the purchaser. And the purple edges are the links. Thered path goes through products and purchased items from the purchaser to

the supplier.

VI. CONCLUSION AND FUTURE WORKIn this paper we refined the definitions of enterprise

relationship network and gave the analysis of thecharacteristics and differences with other social networks.Then we gave more details of ERN model that is composedby ENTG, EMPG, PROG and TLHG. One of the mostimportant features of ERN is to support social business. Wepresented the technique architecture of ERN and describedhow the ERN supports social business. Furthermore, weintroduced a case study on SDCMSP with a simulateddataset. We achieved some relationship visualization andsocial analytics results based on this dataset.Our plans for future work include to continue

investigating the analytics and visualization technologies onERN and to develop suitable means for enterprises toleverage social business to get advantages over thecompetitors. Moreover, with the increasing amount ofenterprises involved in ERN and more applications based onERN being developed, the problems of security and privacyare unavoidable, which will be a valuable topic.

ACKNOWLEDGMENTThe authors would like to acknowledge the support

provided by the National Key Technologies R&D Programof China (2012BAF12B07), the Independent InnovationProjects of Shandong Province (2013CXC30209), and theNatural Science Foundation of Shandong Province(ZR2011FQ031).

358353353353

REFERENCES[1] Liqiang Wang, Shijun Liu, Lei Wu, Pan Li, Xiangxu Meng, "Build

Enterprise Relationship Network to Support Collaborative Business,"Doctoral Symposium of 2014 International Conference onInteroperability for Enterprises Systems and Applications (I-ESA),March 2014

[2] http://en.wikipedia.org/wiki/Enterprise_relationship_management[3] http://en.wikipedia.org/wiki/Enterprise_2.0[4] http://en.wikipedia.org/wiki/Enterprise_social_networking[5] Capuano, N., Gaeta, M., Orciuoli, F., Ritrovato, P., "Semantic Web

Fostering Enterprise 2.0," 2010 International Conference on Complex,Intelligent and Software Intensive Systems (CISIS), pp.1087-1092, 15-18 Feb. 2010

[6] Holroyd, P., "Applications of network theory to social systems,"Physical Science, Measurement and Instrumentation, Managementand Education - Reviews, IEE Proceedings A , vol.130, no.5, pp.273-280, July 1983

[7] Junhua Ding, Cruz, I., ChengCheng Li, "A formal model for buildinga social network," 2011 IEEE International Conference on ServiceOperations, Logistics, and Informatics (SOLI), pp.237-242, 10-12July 2011

[8] Conti, M., Passarella, A., Pezzoni, F., "A model for the generation ofsocial network graphs," 2011 IEEE International Symposium onWorld of Wireless, Mobile and Multimedia Networks (WoWMoM),pp.1-6, 20-24 June 2011

[9] Jin Cao, Hongyu Gao, Li, L.E., Friedman, B., "Enterprise socialnetwork analysis and modeling: A tale of two graphs," 2013Proceedings IEEE INFOCOM, pp.2382-2390, 14-19 April 2013

[10] DiMicco, J.M., Geyer, W., Millen, D.R., Dugan, C., Brownholtz, B.,"People Sensemaking and Relationship Building on an Enterprise

Social Network Site," 42nd Hawaii International Conference onSystem Sciences, 2009. HICSS '09, pp.1-10, 5-8 Jan. 2009

[11] Bennett, S., "The benefits of communicating and collaborating in areal-time enterprise social network," 2012 IEEE International onProfessional Communication Conference (IPCC), pp.1-4, 8-10 Oct.2012

[12] Kohout, J., Neruda, R., "Two-Phase Genetic Algorithm for SocialNetwork Graphs Clustering," 2013 27th International Conference onAdvanced Information Networking and Applications Workshops(WAINA), pp.197-202, 25-28 March 2013

[13] Dong Liu, Li Wang, Jianhua Zheng, Ke Ning, Liang-Jie Zhang,"Influence Analysis Based Expert Finding Model and Its Applicationsin Enterprise Social Network," 2013 IEEE International Conferenceon Services Computing (SCC), pp.368-375, June 28 2013-July 3 2013

[14] Jamali, M., Abolhassani, H., "Different Aspects of Social NetworkAnalysis," International Conference on Web Intelligence, 2006. WI2006. IEEE/WIC/ACM, pp.66-72,18-22 Dec. 2006

[15] Singh, S., Singh, N., "Big Data analytics," 2012 InternationalConference on Communication, Information & ComputingTechnology (ICCICT), pp.1-4, 19-20 Oct. 2012

[16] Jinson Zhang, Mao Lin Huang, "5Ws Model for Big Data Analysisand Visualization," 2013 IEEE 16th International Conference onComputational Science and Engineering (CSE), pp.1021-1028, 3-5Dec. 2013

[17] Bakshi, K., "Considerations for big data: Architecture and approach,"2012 IEEE Aerospace Conference, pp.1-7, 3-10 March 2012

[18] Ido Guy, Inbal Ronen, Eric Wilcox, "Do you know?: recommendingpeople to invite into your social network," Proceedings of the 2009International Conference on Intelligent User Interfaces(IUI’09),pp.77-86, February 8-11, 2009

359354354354