Purpose: To determine the impact of document summarization parameters on the evaluation metrics of Classification algorithms for Persian texts. Methodology: 1000 news texts were collected from yjc. ir news agency website based on the number of visits, with at least 100 and at most 350 words, out of which 250 were selected randomly. Titles, summaries, and the texts of the 250 docs were included in three groups. The number of documents were increased by 100 percent in two stages, to 500 and 1000. After text preprocessing and deleting stop words by programming code, TF-ISF summarization technique was implemented on them. 12 Excel files were created from the words of original texts. Then, Bayesian, Decision tree, SVM and Rulebased algorithms implemented by Rapid Miner software, which provided 120 Excel output files for verifying accuracy, precision, and recall. Finally, five comparisons between the results were considered including comparing of results with 100% increase in the number of documents, comparing the parameters of TF and ISF Summarizer, comparison of Bayesian Classification algorithms, decision tree, Rule and SVM, comparing the original text and summary and comparison of the documents labels. Findings: The results indicated the superiority of evaluation criteria in Classification of 1000 documents relative to those of 250 and 500, which in 84% of cases. Meanwhile, the ISF Summarizer method compared to TF in 82% of comparison showed a greater impact on Classification accuracy. In addition, the values of the accuracy in Bayesian Classification and the SVM were better. The highest value obtained from the accuracy (96. 67%) in the SVM Classification by 1000 documents of original text and ISF Summarizer technique. Conclusion: Appropriate parameters for summarization and efficient Classification techniques can improve the accuracy of Persian text Classification process, while the required time also decreases. The best results obtained in the evaluations show that ISF Summarizer, Bayesian and SVM algorithms, 1000 documents, as well as the main text are more effective.

In most surveys, the occupation and job-industry related questions are asked through open-ended questions, and the coding of this information into thousands of categories is done manually. This is very time consuming and costly. Given the requirement of modernizing the statistical system of countries, it is necessary to use statistical learning methods in official statistics for primary and secondary data analysis. Statistical learning Classification methods are also useful in the process of producing official statistics. The purpose of this article is to code some statistical processes using statistical learning methods and familiarize executive managers about the possibility of using statistical learning methods in the production of official statistics. Two applications of Classification statistical learning methods, including automatic coding of economic activities and open-ended coding of statistical centres questionnaires using four iterative methods, are investigated. The studied methods include duplication, support vector machine (SVM) with multi-level aggregation methods, a combination of the duplication method and SVM, and the nearest neighbour method.

Traditional methods of summarization were very costly and time-consuming. This led to the emergence of automatic methods for text summarization. Extractive summarization is an automatic method for generating summary by identifying the most important sentences of a text. In this paper, two innovative approaches are presented for summarizing the Farsi texts. In these methods, using a combination of deep learning and statistical methods (TFIDF), we cluster the concepts of the text and, based on the importance of the concepts in each sentence, we derive the sentences that have the most conceptual burden. In these methods, we have attempted to address the weaknesses of representation in repetition-based statistical methods by exploiting the unsupervised extraction of association between vocabulary through deep learning. In the first unsupervised method, without using any hand-crafted features, we achieved state-of-the-art results on the Pasokh single-document corpus as compared to the best supervised Farsi methods. In order to have a better understanding of the results, we have evaluated the human summaries generated by the contributing authors of the Pasokh corpus as a measure of the success rate of the proposed methods. In terms of recall, these have achieved favorable results. In the second method, by giving the coefficient of title effect and its increase, the average ROUGE-2 values increased to 0. 4% on the Pasokh single-document corpus compared to the first method and the average ROUGE-1 values increased to 3% on the Khabir news corpus.

This paper proposes a method for processing motor imagery-based Electroencephalography (EEG) signals to generate precise signals for Brain-Computer Interface (BCI) devices used in rehabilitation and physical treatments. BCI research is mainly used in neuroprosthetic applications to help improve disabilities. We analyze EEG data from seven healthy individuals using 59-channel caps. The signals are down-sampled to 100 Hz after pre-processing to remove artifacts and noise by using Filter Bank Common Spatial Patterns (FBCSP). EEG features are extracted using the Fisher Discriminant Ratio (FDR). A comprehensive comparison of Classification methods is conducted, encompassing statistical techniques, machine learning algorithms, and neural network-based models. Specifically, Linear Discriminant Analysis (LDA) and K-Nearest Neighbors (KNN) are evaluated as statistical classifiers; Support Vector Machine (SVM) is used for the machine learning approach; and Radial Basis Function (RBF), Probabilistic Neural Network (PNN), and Extreme Learning Machine (ELM) are explored as neural network models. Model performance is validated using K-fold cross-validation and confusion matrix analysis. Among all evaluated classifiers, the ELM model—implemented as a single-layer neural network—demonstrates superior Classification accuracy, suggesting its strong potential for real-time BCI applications in neurorehabilitation.

Classifying customers using data mining algorithms enables banks to keep old customers loyality while attracting new ones. Using decision tree as a data mining technique, we can optimize customer Classification provided that an appropriate decision tree is selected. In this article we have presented an appropriate model to classify customers who use internet banking service. The model is developed based on CRISP- DM standard and we have used real data of Sina bank's Internet bank. In comparison with the other decision trees, ours is based on both optimization and accuracy factors that recognizes new potential internet banking customers using a three level Classification, which is low/medium and high. This is a practical, documentary-based research. Mining customer rules enables managers to make policies based on discovered patterns in order to have a better understanding of what customers really desire.

Breast cancer is the second major cause of death, and it accounts for 16% of all cancer deaths worldwide. Most of the methods for detecting breast cancer such as mammography are very expensive and difficult to interpret. There are also limitations like cumulative radiation exposure, over-diagnosis, and false positives and negatives in women with a dense breast that pose certain uncertainties in the high-risk populations. The objective of this work is to create a model that detects breast cancer through blood analysis data using the Classification algorithms. This serves as a complement to the expensive methods. High-ranking features are extracted from the dataset. The KNN, SVM, and J48 algorithms are used as the training platform in order to classify 116 instances. Furthermore, the 10-fold cross-validation and holdout procedures are used coupled with changing of random seed. The results obtained show that the KNN algorithm has the highest and best accuracies of 89. 99% and 85. 21% for the cross-validation and holdout procedures, respectively. This is followed by the J48 algorithm with accuracies of 84. 65% and 75. 65% for the two procedures, respectively. The SVM algorithm has the accuracies of 77. 58 and 68. 69%, respectively. Although, it has also been discovered that the blood glucose level is a major determinant in detecting the breast cancer, it has to be combined with other attributes to make decisions as a result of other health issues like diabetes. With the results obtained, women are advised to do regular check-ups including blood analysis to know which blood components are required to be worked on in order to prevent breast cancer based on the model generated in this work.