Knowledge Extraction
Elham Samadi; Hasanali Bakhtiyar Nasrabadi; Zohreh Saadatmand
Abstract
The aim of this research is to employ neural networks in discovering functional knowledge based on the rational training of Avicenna and Kant. The methodology of this study is based on deep learning neural networks, making it an exploratory research. Given the practicality of functional knowledge, this ...
Read More
The aim of this research is to employ neural networks in discovering functional knowledge based on the rational training of Avicenna and Kant. The methodology of this study is based on deep learning neural networks, making it an exploratory research. Given the practicality of functional knowledge, this research is applied in nature. To assess the significance of components and evaluation indicators of functional knowledge, text mining and the frequency of related symbols have been used. In order to utilize data mining techniques in this research, the WEKA software has been employed. The algorithms considered for implementation in this study are MLP, SVR, AdaBoost.R, Bagged Trees (BAGTREE), Linear Regression (LR), and Least Squares Support Vector Regression (LSSVR). According to the results obtained for functional knowledge, the LSSVR and SVR methods outperform the others, indicating their superiority. As the charts illustrate, there is significant volatility in this dataset, making prediction challenging. Furthermore, the R2 value is very close to one, indicating relatively accurate predictions by the methods. Neural networks can serve as powerful tools to aid in rational thinking, logical decision-making, and better understanding of the surrounding world, in line with the perspectives of Avicenna and Kant. These tools can assist in analyzing and interpreting complex data in these fields and strive for rationality and human excellence.
