Variable Assessment of a Start-Up Projects by an Investor Based on Fuzzy Models

The purpose of the research is to develop effective and easy-to-use fuzzy models, which the investor can choose to use as an expert system when assessing the feasibility of investing in a startup project. A set of three fuzzy models for assessing the quality of startups has been developed based on the Mamdani algorithm, which allows the selection of startup projects under conditions of uncertainty of input data. Variability of choosing a certain model from a given set, or consecutive application of two or three models is the prerogative of the investor. Ease of use of these models is an important task that was solved by the authors. An important feature of the basic fuzzy model is the presence in it of an input function of goodwill, the carrier of which is the members of the project team and the value of which is estimated personally by the investor. The investor's subjective assessment of goodwill can be strengthened or, conversely, reduced by taking into account two additional parameters that take into account the accumulated experience of the members of the startup team over all the previous years of their professional activity and the prospects for scaling the startup. These two parameters reflect the characteristics of the startup in quantitative form and do not depend on the degree of investor commitment. Thus, the basic model, which is the basis for the investor to make a decision about the feasibility of financing a startup, takes into account the will of the investor and is built on the principle of personal responsibility for project financing actions. The second fuzzy model is a static model and evaluates a startup based on the predicted duration of its safe development under competitive conditions. By the request of the investor a static model can be added to the examination as an additional risk assessment tool. For this purpose, the authors also developed a third model based on the dynamic parameters of the startup project, which selected the increase in the last quarter of financial and informational indicators. The basic, static, and dynamic fuzzy models for evaluating startup projects are built on three inputs and one output, each contain three membership functions: Gaussian-shaped for inputs, triangular-shaped for outputs. A limited number of input parameters reduces the time of data collection and preparation for modeling, and also allows you to quickly get a project estimate. The simplicity and convenience of the interface allows the investor to personally perform modeling procedures and receive an expert assessment in any option of using a set of fuzzy models. The modeling process is carried out in the MATLAB environment. The use of fuzzy models reduces the subjectivity of expert evaluations, allows establishing the level of investment risk and considers the wishes of the decision-maker. A set of evaluation criteria was formed and simulation results for the “SCHEDRYK-001” startup were given. It was determined that the investment attractiveness of the startup project for the basic model was 52.4 points on a 100-point scale, 41 points for the static model and 25 points for the dynamic model. Data for conducting the research were obtained from the website of the YouControl company (The YouControl system is an online company verification service. https://youcontrol.com.ua/ , last accessed 2023/09/23.), as well as directly at the enterprises of Ukraine. The results of the study make it possible to apply in practice the created expert system for vague evaluation of the investment attractiveness of startup projects.