Using clustering to predict the effectiveness of innovative environmental protection technologies

The article examines the role of innovations in the development of the national economy. The dynamics of budgetary financing of strategic priorities of innovation and technology transfer activities in Ukraine for 2019-2021 is analyzed. The priorities, the support of which is increasing annually, and the priorities where funding is decreasing, are identified. It is proposed to carry out a predictive assessment of innovative projects using the clustering procedure. The implementation of the proposal is highlighted on the example of innovative developments in the strategic priority area "Widespread use of cleaner production technologies and environmental protection". A point system for expert evaluation of the planned scientific results of the innovation and the scientific potential of the performers is shown. The indicators that have the greatest impact on the forecast efficiency are identified. The results of clustering of innovative developments according to certain indicators are presented. The statistical indicators of average scores in each cluster are analyzed. Significant differences between the clusters have been determined using the non-parametric Mann-Whitney test. The carried out clustering provides reasonable recommendations on the feasibility of financing an innovative project based on predictive expert assessments.


Introduction
The basis of socio-economic policy in the modern world is the use of innovative technologies.In Ukraine, no sector of the national economy has a sufficiently high, competitive level of technological development.This problem is especially relevant for the environmental sector.Therefore, state regulation is necessary to increase the level of implementation of environmentally oriented innovations.In particular, Ukraine provides financial support for strategic priorities of innovation and technology transfer activities at the expense of the state budget.
The introduction of the latest technological solutions is preceded by fundamental and applied research.Given the shortage of budgetary funds, it is an urgent task to assess the prospects of certain projects and improve the methods of such assessment.
Strategically important tasks for Ukraine's economy are the development of domestic knowledgeintensive production, development, implementation, and adoption of innovative information techniques and technologies aimed at manufacturing competitive products.The dependence of economic development on innovations has been proven in the works of many scientists 1, 2, 3, 4.Such development requires solving a number of legislative, political, economic, and institutional tasks 5, 6, 7.At the same time, global trends, specifics of the Ukrainian economy, industry characteristics, etc. must be taken into account.Innovations in the use of cleaner production technologies and scoring the planned scientific results of innovative developments and the scientific potential of the performers.Its essence is reflected in the scale of expert assessments (Table 1).This methodology was tested on the example of innovative developments implemented in Ukraine in 2019-2021.The sample included developments aimed at rational subsoil, land and water use, improvement of technologies for cleaning, recycling and reducing waste and harmful emissions.

Initial data
In Ukraine, in 2017-2021, seven strategic priority areas of innovation were funded from the state budget (Table 2).The total budget allocated to strategic priorities of innovation and technology transfer activities in 2021 amounted to 357.7 million UAH.Compared to the previous year, the increase was 7.2% (UAH 333.9 million), and compared to 2019 -25.6% (UAH 268.8 million).Thus, the dynamics of financing strategic priorities in Ukraine in 2019-2021 is positive.
The main focus was on innovations in energy-saving technologies, which grew by 2.6 times in three years.In 2021, compared to the previous year, there was a decrease in the amount of funds allocated for the high-tech development of the transport system, the rocket and space industry, aircraft and shipbuilding, weapons and military equipment (by 4%); the introduction of new technologies and equipment for quality medical care, treatment, and pharmaceuticals (by 64%); and the use of cleaner production and environmental protection technologies (by 29%).Over the past three years, there has been a decline in support for new technologies for the production of materials, their processing and compounding, and the creation of the nanomaterials and nanotechnology industry.Funding for medical innovations has been halved.
In 2021, the strategic priority "Widespread use of cleaner production and environmental protection technologies" ranks fifth in terms of funding.The amount of budgetary funding for innovation activities under this strategic priority in 2021 amounted to UAH 19182.86 thousand, or 5.4% of the total budgetary funding for strategic priorities.Compared to 2020 (UAH 27051.17thousand or 8.1%), the amount of funding decreased by 1.4 times, and its share in total funding also decreased, which moved the priority from fourth position in 2020 to fifth in 2021.
In order to improve the forecast of the effectiveness of research and development work on the development of innovative projects at the planning stage, we believe that it is advisable to use fuzzy logic to ensure proper "intellectual" analysis.
The implementation of the recommendations will be illustrated by the example of the evaluation of innovations carried out in 2019-2021 within the framework of financing the technology for cleaner production and environmental protection.Such developments include medium-term projects (over 50%), as well as long-term (up to 30%) and short-term (up to 20%) projects.
Five medium-term priorities were approved under this strategic priority: 1. Application of technologies for rational subsoil and land use.2. Implementation of advanced technologies for water supply, water use and water disposal.3. Application of closed-cycle technologies, technologies for cleaning, processing and utilization of industrial and household waste.4. Application of radioactive waste management technologies and reduction of their negative impact on the environment.5. Application of technologies to reduce harmful emissions and discharges.
The dynamics of funding for medium-term projects in the area of "Widespread use of cleaner production and environmental protection technologies" is shown in Table 3.In 2021, the amount of funding amounted to more than UAH 11.7 million, or 61% (2020 -49%, 2019 -33%) of the strategic priority funding.During 2019-2021, the vast majority of funds (UAH 8188.3 thousand or 70% in 2021) were allocated for the application of technologies for rational subsoil and land use.The share of this area in financing in 2020 was 70.4%, and in 2019 -81.6%.The smallest amount of funds was allocated for the implementation of innovations in the use of radioactive waste management technologies and reduction of their negative impact on the environment (UAH 500.0thousand or 4.3%).In 2021, funds were allocated by type of innovation activity as follows: "Purchase of machinery, equipment and software" -UAH 602.8 thousand (5%); "Creation and development of innovative infrastructure" -UAH 173.1 thousand (2%); "Other" -UAH 10941.9thousand (93%).

Discussion of experimental results
In total, in 2019-2021, more than 100 medium-term innovative developments were implemented in Ukraine in the area of "Widespread use of cleaner production and environmental protection technologies."Based on the proposed approach, the expert group evaluated 52 innovative developments.The expert group consisted of three members: a representative of academia, an industrialist, and an environmental specialist.
It turned out that the following four indicators are the most important for determining the degree of innovation efficiency: novelty, expected effect from implementation, methodological level and material and technical base of the research, and qualifications of the main researchers.
As a result of dividing the average scores obtained during the examination of innovations by the established minimum acceptable level, the relative values of the assessment of a particular innovation were calculated.
Subsequently, it was necessary to group the obtained estimates.The best method for this is clustering.Clustering has been widely used in studies related to environmental protection.Researchers used different clustering methods.For example, the k-means and hierarchical agglomerative methods were used in studies of air pollution 14.
In general, the analysis of possible clustering methods is considered in 15, where a simple implementation of the K-means algorithm was analyzed and a simulation example was performed.Research shows that this algorithm has strong universality and can be applied to most data sources, providing a theoretical basis for searching and analyzing big data.
To solve the problem of fuzzy clustering, we used the Fuzzy Clustering and Data Analysis Toolbox 16.As a result, three clusters were formed: the first one included 14 innovative developments, the second -23 developments, and the third -15 developments.
In the next step, the sum of points for each innovative development in the three clusters was calculated according to the indicators that allow assessing the environmental status.Points were awarded for a particular innovation's positive impact in applying cleaner production and environmental technologies.Statistical characteristics of the average scores in each cluster are presented in Table 4. Comparison of the total scores obtained in different clusters using the non-parametric Mann-Whitney test 17 for the level of significance (P<0.01)indicates that there are significant differences between the first and second cluster (10.1±0.6), the first and third cluster (7.2±0.5), and the second and third cluster (5.2±0.6).
Innovative developments that are included in the first cluster based on expert assessments received the highest number of points in the assessment of their effectiveness.The range of variation in this cluster is also the largest.It should also be noted that the minimum and maximum scores between the clusters differ quite significantly.In the first two clusters, the dispersion of scores is higher than in the third, which indicates a high similarity of innovations attributed to this cluster.
An analysis of the structure of indicators in the clusters revealed that the third cluster includes ineffective innovations, as their implementation did not produce significant environmental results.The first cluster includes the most effective innovations, and the second -intermediate ones.

Conclusions
The proposed methodological approach to predicting the effectiveness of innovative developments involves clustering based on an expert assessment of the planned scientific results and the scientific potential of the performers.Out of 52 innovative developments in the area of "Widespread use of cleaner production and environmental protection technologies", 14 are estimated to be effective.Twenty-three developments require additional evaluation, and 15 raise doubts about their effectiveness.The clustering of the same developments by actual results confirmed the adequacy of the forecast estimates.All 14 developments included in the first cluster according to the forecast estimates provided a tangible environmental effect as a result of their implementation.Only 10 developments proved to be ineffective.At the same time, 9 of them belong to the third cluster according to the forecast estimates.
The assignment of an innovation to a particular cluster based on expert assessments of the selected informative indicators allows for a sufficiently accurate prediction of the possible effectiveness of the innovation.Accordingly, a decision on financing an innovation that is included in the first or third cluster can be made immediately (accept or reject).If an innovation included in the second cluster is being considered, additional factors need to be analyzed to make a final decision.

Table 1 .
A system for scoring the planned scientific results of innovative development and the scientific potential of performers

Table 2 .
Dynamics of budget financing of innovations by priority areas

Table 3 .
Financing of medium-term priority areas of innovation and technology transfer under the strategic priority area "Widespread use of cleaner production and environmental protection technologies"

Table 4 .
Cluster distribution of the average values of the efficiency scores of innovative developments by cluster