1st ICAI 2020

International Conference on Automotive Industry 2020

Mladá Boleslav, Czech Republic

First of all, using descriptive statistics, the enterprises included in clusters 1 and 2 were analyzed. The analysis was conducted primarily for structural variables. The results of the analysis showed an asymmetrical distribution of values of structural variables (right-hand asymmetry, bevel distribution towards lower values). A similar picture is characteristic of the distribution of values of structural variables in clusters 2 and 3. For Cluster 5 such studies have not been conducted, as it includes only one enterprise. Due to size limitations of this article, it is not possible to provide detailed information on all structural variables for the analyzed clusters. Therefore, below will provide information only on variable X1 (number of units) in clusters 1 and 4 (see Table 5).

Table 5: Analysis of distribution of variable X1 (number of divisions) in clusters 1 and 4

Value of the variable “number of divisions” in the cluster 1

Value of the variable “number of divisions” in the cluster 4

Indicator name

Mean value

1.6 1.0 1.0

1.5 1.0 1.0

Mediana Fashion At least .

1 7

1 5

Maximum .

Source: own calculations

In the cluster 1 out of 270 enterprises 167 (61.9%) have one unit. Another 61 enterprises (22.6%) have two divisions and 28 enterprises (10.4%) have three divisions. The remaining 14 enterprises (5.1%) have 3 to 7 subdivisions. In the cluster 4 out of 10 enterprises 8 (80%) have one unit. One enterprise (10%) has two divisions and one enterprise (10%) has five divisions. At the same time, if we analyze the value of variable X4 (total assets) for two enterprises with several subdivisions, it will be significantly lower than the average value of this variable for this cluster. Thus, it can be concluded that Cluster 1 mainly gathers both small production enterprises / workshops and networks of small production enterprises / workshops. At the same time, the vast majority of enterprises in this cluster are small production enterprises / workshops. In Cluster 4, industrial enterprises are predominantly assembled. The remaining two enterprises, judging by the size of assets (X4), are a network of small production enterprises / workshops, which were included in Cluster 4 due to their higher efficiency compared to the enterprises of Cluster 1. If the enterprises included in cluster 4 are quite easy to conduct due to its small size, then cluster 1 requires additional research. For this purpose, the authors decided to build a decision tree for the variable number of jobs (X1) in order to identify the key factors affecting its value. The results of building the decision tree are shown in Figure 1. The method of a decision tree gives the chance to reveal certain groups of the enterprises from the point of view of the parameters influencing quantity of workplaces. This method is less demanding to distribution parameters, than classical statistical methods of the data

25