الگوریتم ایجاد یک مدل از ترکیب بهینه پنبه
Abstract
1-Introduction
2-Data Preprocessing
3-Data Feature Extraction
4-Model Establishment
5-Model Evaluation
6-Model Calibration
7-Experiment and Simulation
8-Conclusion
Acknowledgement
References
Abstract
Due to the small amount of sample data and poor quality in cotton mills, this leads to large errors in the prediction of yarn quality to make the model difficult to train. This paper uses the appropriate data stretching algorithm after data preprocessing. Then, based on the features extracted by the main factors, the mapping relationship model of cotton blending is established. Finally, the evaluation and the correction of model are established. Actual tests have shown that the cost of cotton mills is decreased about 15% by the system and is increased about 25% for the profit by the system.
Introduction
Chuang Gao [6] studied the prediction model of yarn quality based on the combination of genetic algorithm, principal component analysis and competitive neural network. Yang S [7] established a physical or theoretical model to describe the relationship between yarn structure and yarn properties. And they provided the data needed for the model. Kothari N [8] showed that the degree of variation in plant fiber length and length uniformity were related to genotype. Chattopadhyay R [9] discussed how to use artificial neural networks to predict yarn properties from fiber parameters. Muhlstadt M [10] studied the fiber volume fraction gradient (G) in the laminated plain woven fabric (PWF), and solved the shortening problem by appropriately measuring the G in the laminated PWFs, and proposed the corresponding model. Ji R [11] proposed a cotton heterogeneous fiber classifier based on support vector machine (decision tree support vector machine, DTSVM). Experiments show that the recognition rate of different heterogeneous fibers is greater than 92%. Dhawan S [12] used the binary classification concept of support vector machine and the multi-class classifier concept of neural network to identify the spam and data categories received online and achieved good results.