بکارگیری تکنیک های محاسبات تکاملی برای شناسایی نوآوران در جوامع نوآوری آزاد
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|2390||2013||8 صفحه PDF||سفارش دهید||5960 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Expert Systems with Applications, Volume 40, Issue 7, 1 June 2013, Pages 2503–2510
Open innovation represents an emergent paradigm by which organizations make use of internal and external resources to drive their innovation processes. The growth of information and communication technologies has facilitated a direct contact with customers and users, which can be organized as open innovation communities through Internet. The main drawback of this scheme is the huge amount of information generated by users, which can negatively affect the correct identification of potentially applicable ideas. This paper proposes the use of evolutionary computation techniques for the identification of innovators, that is, those users with the ability of generating attractive and applicable ideas for the organization. For this purpose, several characteristics related to the participation activity of users though open innovation communities have been collected and combined in the form of discriminant functions to maximize their correct classification. The right classification of innovators can be used to improve the ideas evaluation process carried out by the organization innovation team. Besides, obtained results can also be used to test lead user theory and to measure to what extent lead users are aligned with the organization strategic innovation policies.
The concept of open innovation, launched by Chesbrough (2003) and others, has become increasingly popular among scholars and industry practitioners since the term was coined. Open innovation refers to the use of external sources and actors to achieve innovation, and it is based on the idea that companies should not just rely on internally developed ideas and knowledge, but increasingly also on ideas and knowledge developed externally (Chesbrough et al., 2006 and Tödtling et al., 2011). It assumes that useful knowledge is widely diffused and abundant. There is, for example, a growing availability of knowledge from multiple innovation actors, including universities, specialized suppliers, inventors and knowledge brokers. In this conditions, the “old” model of closed innovation where innovation processes are controlled by the company needs to be changed in favor of the detection and assimilation of externally developed knowledge (Barge-Gil, 2010, De Jong et al., 2010 and Poetz and Schreier, 2012). Previous studies agree that open innovation is not a general phenomena and depends on certain company characteristics as well as external conditions. Chesbrough (2003) identified various external factors that explain why enterprises increasingly adopt the open paradigm. The availability of a strong public knowledge base, a mobile and educated working population or the availability of ample external finance for innovation are the three conditions enabling open innovation to emerge. From the viewpoint of the organization, there are various mechanisms and channels used for sourcing and acquiring external knowledge such as the absorption of local knowledge spillovers, collaboration in R&D and innovation with firms and universities, relations to spin-off companies, informal knowledge interactions, customer contributions through design toolkits or idea competitions (Keeble and Wilkinson, 2000, Schwab et al., 2011, Tödtling et al., 2006 and Von Hippel and Katz, 2002). The strategic challenge is how firms can best organize the sourcing, codification and exploitation of the internal and external knowledge and informational resources to maximize and sustain innovation (Love & Roper, 2009). One of the most popular mechanism for open innovation implementation is user innovation communities (Dahlander, Frederiksen, & Rullani, 2008). Firms such as Microsoft, Dell, IBM, BMW, and Nokia increasingly invest in virtual communities to solicit user contributions as part of their innovation processes. This trend is explained by the increase in digitalization and the decrease in the costs of communication that have lead to an exponential growth of user innovation platforms (Mahr & Lievens, 2012). Internet have facilitated the accessibility of these platforms by users geographically distributed all over the world. However, this accessibility is also causing the generation of a huge amount of information which it is difficult to process and evaluate by the innovation departments or experts within organizations. Posted ideas must be evaluated one by one by the innovation department or even some specific experts of the organization, and this evaluation consists of reading the idea, assessing its applicability attending to the strategic innovation policies of the organization and planning their possible implementation in case they are finally accepted. The problem is that online user innovation communities can generate hundreds or even thousands of solutions in a short period of time, saturating the capacity of internal evaluators and hiding the really attractive innovations. That is the reason why online user innovation communities typically include some type of scoring systems so that the community can evaluate potential solutions. This scoring scheme is based on the idea that the crowd can do a better evaluation than individuals since they own a group-based intelligence which can outperform individual knowledge (Surowiecki, 2004). However, strategic innovation policies of organizations are not always aligned with users’ desires. Some non affordable ideas can be excellent for users but prohibitive for the company, and these ideas would probably receive a high score by other community users. In this sense, it is much more useful for the company the identification of users posting ideas that will be finally adopted. This information can be easily collected from innovation communities websites as they usually inform users about the status of their posted ideas. The purpose of this paper consists of the identification of these innovators, defined as those users generating ideas that will be finally adopted by the company. The condition of being innovator or non innovator is a dichotomic property of each user. Therefore, the identification of innovators is a classification problem that can be solved using a discriminant function over a set of variables characterizing the activity and behavior of users within the community, which on the other hand is the main available information. The main problem associated to this identification is that the considered dependent variable contains a high number of zeros (non-innovators), leading to the so called zero inflated problem. To solve this issue, a optimization procedure consisting of finding the values of the variables coefficients so that the discriminant function can maximize the percentage of correct classification of innovators and non innovators is formulated. Three different evolutionary computation techniques are used to solve the problem for evaluating the reliability of results. Additionally, a bootstrapping technique has also been implemented to obtain the confidence intervals of the resulting coefficients. The rest of the paper is structured as follows. Section 2 details previous works related to the open innovation paradigm and the identification of users with special profiles. Section 3 describes the formulation of the problem in the form of an optimization problem and presents the three proposed evolutionary computation techniques: simulated annealing, particle swarm optimization and genetic algorithms. The three algorithms are then applied to the case study of IdeaStorm website, which is introduced in Section 4, as well as those variables measuring the activity and behavior of users within this innovation community. Obtained results are discussed in Section 5. Finally, conclusions are provided in Section 6.
نتیجه گیری انگلیسی
This paper deals with the problem of identifying innovators in open innovation communities using variables related to their activity. Mathematically, innovators are estimated using discriminant functions obtained by a linear combination of the selected variables. Due to the zero inflated characteristic of the dependent variable, the problem has been formulated as an optimization problem consisting of determining the coefficient values of the discriminant function so that the innovators and non innovators identification ratios are maximized. Three different optimization techniques have been used for this purpose in order to validate the results. Each algorithm was executed 25 times to average the coefficients’ values and a bootstrapping technique was then applied to obtain the 95% confidence interval. From the viewpoint of the methodology, obtained results show that GA and PSO solve the optimization problem better than SA, leading to best results in terms of classification as well as in terms of smaller ranges in the confidence intervals. From the viewpoint of the application problem, obtained results clearly show that the interactions among users through comments (both sent and received) are better indicators of innovative profiles than the interactions though the scoring system.