عوامل مؤثر بر ارزش دارایی های نامشهود : روش های داده کاوی

It is very important for investors and creditors to understand the critical factors affecting a firm’s value before making decisions about investments and loans. Since the knowledge-based economy has evolved, the method for creating firm value has transferred from traditional physical assets to intangible knowledge. Therefore, valuation of intangible assets has become a widespread topic of interest in the future of the economy. This study takes advantage of feature selection, an important data-preprocessing step in data mining, to identify important and representative factors affecting intangible assets. Particularly, five feature selection methods are considered, which include principal component analysis (PCA), stepwise regression (STEPWISE), decision trees (DT), association rules (AR), and genetic algorithms (GA). In addition, multi-layer perceptron (MLP) neural networks are used as the prediction model in order to understand which features selected from these five methods can allow the prediction model to perform best. Based on the chosen dataset containing 61 variables, the experimental result shows that combining the results from multiple feature selection methods performs the best. GA ∩ STEPWISE, DT ∪ PCA, and the DT single feature selection method generate approximately 75% prediction accuracy, which select 26, 22, and 7 variables respectively.

The market value of a firm’s shares ultimately reflects the value of all its net assets. In the industrial era, physical assets, such as land, capital, and labor were critical factors to judge a firm’s value. However, in modern economies, the development of communication technology, electronic commerce, and the Internet enables these resources to circulate around the world quickly, letting the knowledge-based economy era evolve. Therefore, the important factors for successful companies are the capability and the efficiency in creation, expansion, and application of knowledge [27]. The method for creating firm value transfers from traditional physical production factors to intangible knowledge. In this situation, a large part of a firm’s value may reflect its intangible assets. To evaluate the firm’s value, we not only consider the tangible assets, but also respect the power of intangible assets [9] and [14]. Intangible assets are a firm’s dynamic capability created by core competence and knowledge resources, including organization structure, employee expert skills, employment centripetal force, R&D innovation capability, customer size, recognizable brand, and market share. Many studies [18], [19], [32] and [59] investigate various types of important factors in intangible assets valuation. Gleason and Klock [19] indicate that advertising and R&D expenditures are positively related to Tobin’s Q, a proxy for intangible assets, but the firm size has a negative relation with it. Wiwattanakantang [59] examines the effect of controlling shareholders, including various types of controlling, participation in management, and pyramids on corporate value and finds no significant effect on Tobin’s Q. Fukui and Ushijima [18] investigate the industry diversification of the largest Japanese manufacturers. Regression results show that the average relationship between diversification and intangible assets is negative. However, research to date [36] and [59] provides mixed evidence on the various factors affecting intangible assets. Regarding literature, studies focusing on different domain problems discover different factors that affect intangible assets. This leads to an important research question of what factors are more representative to judge or even predict a firm’s intangible assets value. Therefore, this study first reviews related literature from diverse domains including accounting, finance, management, and marketing to collect relatively important factors affecting intangible assets. Then, we consider feature selection to select important features (or factors) from a given dataset. In data mining, feature selection is a very important step for obtaining quality mining results [21], as it aims to filter out redundant or irrelevant features from the original data [63]. The remaining selected features are more representative and have more discriminative power over a given dataset. There are a number of statistical and data mining based feature selection methods that are widely used for many business domains [8], [23] and [47]. In this study, we consider five feature selection methods to assess their prediction performance of intangible assets. They are principal component analysis, stepwise regression, decision trees, association rules, and genetic algorithms. Because of fewer regulations and less disclosure of intangible capital, financial reporting cannot actually reflect the value of intangible assets. The problem in the traditional financial accounting framework is that reporting lacks the recognition of intangible capital value and creates an information gap between insiders and outsiders [58]. Therefore, we expect the empirical results in this paper to allow us to not only understand the best feature selection method for effectively evaluating intangible assets but also to provide other information different from financial statements. This will help investors and creditors to better evaluate the investment or lending opportunities, and help them make more effective decisions. In other words, the contribution of this paper is two-fold. First, for the feature selection field, this paper shows the applicability of the state-of-the-art feature selection algorithms to the domain of intangible assets evaluation, which has never been studied before. Second, for intangible assets evaluation, this paper attempts to use the state-of-the-art feature selection algorithms to identify critical factors affecting a firm’s value. The remainder of this paper is organized as follows: Section 2 reviews related studies about firm value and also briefly describes the feature selection methods; Section 3 describes the experimental methodology; Experimental results are presented in Section 4. Finally, the conclusion is provided in Section 5.

As the method for creating firm value transfers from traditional physical assets to intangible knowledge, it is commonly found that the market values of knowledge-based firms are much higher than the book values. Therefore, valuation of intangible assets becomes a widespread topic of interest in the new economy. In order to effectively evaluate intangible assets, this study employs data mining techniques to identify important factors affecting intangible assets. Particularly, feature selection, the data pre-processing step of data mining, is considered to select and extract more useful information in the massive related materials. This paper uses five feature selection methods, which are principal component analysis (PCA), stepwise regression (STEPWISE), decision trees (DT), association rules (AR), and genetic algorithms (GA). To assess the effectiveness of the identified features of these methods, multi-layer perceptron (MLP) neural networks are constructed as the prediction model to examine the prediction performances. Regarding the experimental results over the chosen dataset containing 61 variables, for single feature selection methods, DT is the best method to provide the highest rate of prediction accuracy and lowest rate of Type I errors. In addition, it only selects seven features, which constructs the prediction model in a very efficient manner. On the other hand, for the multiple feature selection methods created by combining the results from different single methods, GA ∩ STEPWISE and DT ∪ PCA are the top two methods. They select 26 and 22 features respectively and provide the best prediction accuracy and outperform DT. Therefore, these selected features can then be regarded as the important factors affecting intangible assets. In addition, these findings can allow investors and creditors to make decisions about investments and loans. It should be noted that although this paper considers several widely used feature selection methods, there are other related algorithms available in literature, such as genetic algorithms [15], rough set [11]. However, from the practical standpoint, it is difficult to conduct a comprehensive study on all existing feature selection methods. In addition, currently it is hard to define the most representative method in the intangible assets evaluation domain, and there is not a comparative study based on these methods. The absence of such a study can be regarded as one of the future research issues. Moreover, since this study mainly focuses on evaluating the intangible assets and market-based value problems, other domain problems can be applied in future work, such as financial distress forecasting in the financial domain, which includes audit opinion prediction, auditor’s going concern uncertainty decision prediction, and litigation prediction in account/auditing domain, etc. In other words, when the best feature selection methods for various business domain problems are identified, useful managerial implications can be derived from the selected features.