Due to the influence and limitations of the multisourced, heterogeneous, and unbalanced characteristics of embedded multifunctional data, the application effect of the current data mining technology is not good, and the accuracy is low. To solve the above problems, an embedded multifunctional data mining technology based on granular computing was studied. According to the three characteristics of embedded multifunctional data, preprocessing such as data reduction, data standardization, and data balance were implemented. We implemented data granulation for the preprocessed data and calculated the data granulation characteristics, including offset, particle density, and intraparticle interval. Taking granular features as the input content, embedded multifunctional data mining was realized by using a neural network to complete the objectives of data classification, anomaly detection, fault identification, and so on. The experimental results showed that the anomaly mining results of each type of data mining were greater than 0.9, indicating that the accuracy of the mining technology is high.

Aiming at the problems of premature convergence of existing workshop dynamic data scheduling methods and the decline in product output, a flexible industrial job shop dynamic data scheduling method based on digital twin technology is proposed. First, digital twin technology is proposed, which provides a design and theoretical basis for the simulation tour of a flexible industrial job shop, building the all-factor digital information fusion model of a flexible industrial workshop to comprehensively control the all-factor digital information of the workshops. A CGA algorithm is proposed by introducing the cloud model. The algorithm is used to solve the model, and the chaotic particle swarm optimization algorithm is used to maintain the particle diversity to complete the dynamic data scheduling of a flexible industrial job shop. The experimental results show that the designed method can complete the coordinated scheduling among multiple production lines in the least amount of time.

The purpose of image smoothing is to improve the visual effect of the image and improve the clarity of the image, so as to make the image more conducive to computer processing and various feature analysis. Because the current technology fails to smooth the preprocessed image, it leads to the extraction of image texture features. The anti-interference performance is weak. For this reason, an image texture feature extraction technology based on the digital twin is proposed. Similarity analysis is carried out through the internal structure of the image, and the image is smoothed by the semisupervised learning method. On the basis of optimizing the denoised image through digital twinning, detect target feature points in the original image, then remove the abnormal and split feature points, assign the direction of image texture feature points, and build a fuzzy back propagation neural network model. Image texture feature extraction technology is implemented. The experimental results show that, compared with the traditional method, the proposed technique has a strong identification of original image features, and has a strong consistency with original data, and has a strong ability to resist the influence of abnormal data, noise, or redundant feature points.

Aiming at the problems of low accuracy, the long time required, and the large memory consumption of traditional data mining methods, a local discrete text data mining method in high-dimensional data space is proposed. First of all, through the data preparation and preprocessing step, we obtain the minimum data divergence and maximize the data dimension to meet the demand for data in high-dimensional space; second, we use the information gain method to mine the pre-processed discrete text data to establish an objective function to obtain the highest information gain; finally, the objective functions established in data preparation, preprocessing, and mining are combined to form a multi-objective optimization problem to realize local discrete text data mining. The simulation experiment results show that our method effectively reduces the time and improves the accuracy of data mining, where it also consumes less memory, indicating that the multi-objective optimization method can effectively solve multiple problems and effectively improve the data mining effect.