Comparison of the Regression Method and The Neural Network Method in Specific Cases of Engineering Practice

Comparison of the Regression Method and The Neural Network Method in Specific Cases of Engineering Practice

Blog Article

The presented article presents a summary of research findings based on a comparison of two statistical methods, both of which were applied to the same problem.These were data that were measured during the biomass compaction process, with softwood used as the input material.The aim of the analysis was to create models for both selected statistical methods that would sufficiently describe the given process and subsequently compare them.The second very important step was to compare the applicability of these methods toor knives serpent from the point of view of limiting criteria and to define their advantages and disadvantages in practical use.

For the purposes of the experiment itself, the biomass pressing process was used using a uniaxial press.In the case of this process, parameters were defined that significantly affect the quality of the resulting press, which is represented by the measured quantity - the press density.The parameters already mentioned, which by their nature significantly affect the quality of the press, are the pressing pressure, pressing temperature, humidity of the pressed material and the size of the pressed material fraction.The process itself and the subsequent experimental obtaining of results from the given research took place in laboratory conditions using an experimental uniaxial press.

Subsequently, through the steps of statistical analysis, we carried out estimates of effects, individual tests of hypotheses about the significance of the model and effects and, last but not least, also the predictive ability of the models thus obtained.The obtained models, which were created using the two methods already mentioned, were subsequently compared based on their predictive ability, specifically through the so-called predictive ability of the model, which is represented by the coefficient of determination R2, which is defined as the ratio of the sum of squares SSM explained by the model to the total sum of squares SST.As a result, cleveland browns scarf it expresses the degree of agreement of the observed values with the model.Based on the experimental knowledge thus obtained, the aim is to point out the significance and importance of statistical methods and methodologies that are useful in processing data of various nature and scope.

It is precisely the modelling of a process, which is often complex in nature or demanding on the data obtained, that needs to be differentiated in the statistical methods used.The research results presented in the paper demonstrate their significance.