Abstract
Without powerful analytical tools, people will not be able to uncover the wealth of information hidden behind the data. In order to strip the information from the massive data, it is urgent to find a reasonable solution to deal with the challenge of information overload, and to develop an intelligent and automated alternative tool. Data mining technology transforms the original data into a form suitable for operation by integrating and learning data, extracting useful data for mining, and applying various strategies of data mining to generate useful patterns and rules. Based on this, people can make predictions on new sample data and obtain information. This paper analyzes and describes the problems faced by the inheritance of Chinese folk painting art, including the lack of knowledge of the inheritors, the single inheritance content, and the crisis of the inheritance method. In view of the problems faced by the inheritance of folk painting art, this paper proposes to inherit Chinese folk painting art from the aspects of value edification, rich content, and multiple forms. We analyze the different inheritance methods of contemporary national cultural inheritance and point out that different inheritance views have a certain influence on the formulation of cultural inheritance and the choice of inheritance methods. The preprocessing part of the original data is improved before the training on the data starts. By normalizing the order of magnitude of the original data, the values of the data are all in the same order of magnitude. Taking the least mean square algorithm and the BP neural network classification algorithm as the basic algorithm idea, using the rules of batch learning, and adding the momentum term factor, an improved batch learning BP algorithm is designed. The experimental results show that the improved batch learning BP algorithm can better solve the classification problem. Chinese folk painting art can satisfy people's spiritual and material needs, and also has a positive effect on promoting people's physical and mental development.
1. Introduction
In the context of cultural globalization, the inheritance and protection of Chinese folk painting art has become a consensus [1–3]. Nonetheless, it also faces challenges brought about by the development of modern civilization. In the process of rapid development of modern civilization and in the face of huge economic temptations, some human beings have gradually lost their awareness of the protection and rational development and utilization of traditional cultural resources, including Thangka and other Regong artworks. Nowadays, with the rapid development of the market economy, many difficulties have arisen, among which the most typical problems include the extreme shortage of artistic personnel and the shoddy production of products, the loss of unique skills, and the lack of creative standards [4]. These problems have led to the shortage of fine quality folk painting in China, a crisis with no successor. These situations show that under the impact of the globalization of modern civilization, we urgently need to solve an important problem: how to protect and scientifically and rationally develop national cultural resources in a good and orderly manner.
In China, in addition to Chinese painting, the most systematic, complete, and distinctive painting is Tibetan painting. Thangka is one of the main forms of Tibetan painting. The development of its theoretical research is conducive to promoting the development of Chinese art theory [5]. To study the inheritance of Chinese folk painting art is to protect and rescue Chinese folk painting art. It is also an effective way to cultivate practical ability and theoretical foundation of traditional folk arts and crafts and be able to engage in Thangka painting production. The development of painting art urgently needs this meaningful attempt. In today's society, with the gradual and in-depth development of reform and opening up, Chinese folk painting art has also been impacted by the social environment and humanistic thoughts, so it has gradually developed in the direction of diversification and individualization [6]. Compared with the general inland culture, the origin and development of Chinese folk painting art also has regional characteristics. It is not only the religious ideology of the Tibetan people, but also the carrier for the Tibetan people to communicate with heaven and Earth. It contains the cultural information of human aesthetic concepts, ways of thinking, religious beliefs, and customs in Tibetan areas. Therefore, we can say that the research and exploration of Chinese folk painting art affect the inheritance of thangka to a certain extent.
With the development of the times, especially since the beginning of the 21st century, the art of thangka continues to realize its own inheritance in a new way. Based on the spiritual connotation of the Tibetan nation and the oriental cultural outlook, and on the premise that thangka art inherits the tradition, proceeding from academic thinking, artistic rules, and social life, and chooses nationality, modernity, and regionality, it reflects the uniqueness of the Tibetan nation. In today's society, with the evolution of thangka consciousness and the popularity of inheritance, thangka art has become the research object of folklore, ethnology, and society in many aspects, as well as the research object of aesthetics and fine arts. Therefore, summarizing the inheritance and development of thangka art and studying how to enrich Tibetan art also play a positive role in promoting it.
This paper discusses how to effectively preprocess the original data before data mining. The improvement strategy is to normalize the order of magnitude of the data when the data is input, so as to improve the data. The effect of preprocessing directly affects the effect and accuracy of the subsequent data mining process, and the experimental data is compared and explained, which verifies that the improved method in this paper is effective. On the basis of studying the theory of batch learning BP algorithm, this paper gives an effective improvement method and gives a detailed description of the relevant implementation methods and applications. The momentum term factor is introduced, and the improved batch learning BP algorithm is realized. At the same time, the effectiveness of the improved algorithm in this paper is verified by the comparison of the experimental results and data. As a complex of folk art and religious art, Tibetan Chinese folk painting art has always represented the silhouette of Tibetan cultural life. The previous research results and art monographs tend to be similar to the research on this artistic style, and there are few studies on the inheritance crisis and countermeasures of Chinese folk painting art. This paper analyzes the crisis faced by the inheritance of Chinese folk painting art, analyzes the educational factors of the inheritance of Chinese folk painting art, and finally puts forward its own opinions on the inheritance method. In this subject, due to the limitation of the author's knowledge and scientific research ability, there are still many imperfections. For example, the theoretical research is not deep enough, the empirical materials are relatively rough, and the effectiveness and feasibility of the proposed countermeasures have yet to be tested in practice.
2. Related Work
A predictive model is the one in which an outcome can be determined precisely based on the value of a data item. The data used to mine predictive models is also clearly known. The descriptive schema is a description of the rules that exist in the data [7].
The purpose of data summarization is to compress and extract the data and give a concise and general description of it. The traditional method of data summarization is to calculate the sum, count, average, variance, and other statistical values on each field of the database, or to represent it with graphical methods such as histograms and pie charts [8]. Data mining hopes to discuss data summarization from the perspective of data generalization. Because the information contained in the data in the database is always the most primitive and basic information, sometimes it is necessary to process or browse the data from a higher-level view, so the data should be generalized at different levels to meet various query requirements [9]. There are two main techniques for data generalization: multi-dimensional data analysis methods and attribute-oriented induction methods.
Users often need abstract meaningful descriptions. Inductive abstract descriptions can summarize a large amount of information about a class. There are two typical descriptions: feature description and discriminant description. The feature description is to extract the feature expressions about the data from a set of data related to the learning task, and these feature expressions express the overall characteristics of the data set, while the discriminant description describes the difference between two or more classes. The purpose of classification is to propose a classification function or classification model (also known as a classifier), which can map data items in the database to one of the information categories. Both classification and regression can be used for prediction, and classification outputs discrete category values. The construction methods of classifiers include statistics, machine learning, and neural networks. The effect of classification is generally related to the characteristics of the data. At present, there is no classification method that can adapt to data with various characteristics.
Relevant scholars advocate the development of arts and crafts as one of the representative scholars of cultural undertakings, and propose to coordinate the contradiction between the cultural undertakings of arts and crafts and the economic industry and develop together [10]. They believe that traditional arts and crafts are both material and immaterial. In modern society, traditional arts and crafts have industrial value, employment value, market value, and cultural value. But, at the same time, it faces three major contradictions: economic industry and cultural undertakings, technological renewal and protection of intangible cultural heritage, and cultural changes and inheritance of traditional arts and crafts [11]. It is suggested that the traditional craft needs to be solved by the government and relevant departments in coordination, and it must be implemented in terms of organization, system, personnel arrangement, and funds. At the same time, he also pointed out that “in the era of modern machine industry, it is necessary to further develop another function of arts and crafts, that is, to produce daily necessities suitable for modern new life based on the characteristics of manual single-piece production [12].”
Relevant scholars believe that arts and crafts cannot be industrialized [13]. They emphasized that in the transitional period of modern times, traditional handicrafts and arts and crafts are bound to face many specific problems and need to be solved [14]. However, we need to take it seriously when dealing with the basic concepts of “crafts and fine arts”, “tradition and innovation”, and “culture and industry”. Relevant scholars do not deny the fact that traditional handicrafts do not necessarily meet the needs of modern life, but there is no need to cut their feet to fit their feet [15]. He believes that excessive innovation may lose the significance of handicrafts, especially the impact of traditional handicrafts. Traditional handicrafts are unique and differentiated and are handmade emotional items; while industrial production is characterized by standardized, integrated, horizontal, and vertical. It is a mechanical product [16, 17]. From its characteristics and production requirements, the two are completely contradictory. Therefore, industrialization is only applicable to industries that meet the conditions of industrialization. It is not against the industrialization of the arts and crafts industry, but against the full industrialization. Therefore, relevant scholars put forward that “as far as arts and crafts enterprises are concerned, the popular products in the middle part can be industrialized, but the two ends must be broken; that is, the production of fine arts and crafts and the production of folk handicrafts cannot be industrialized [18–20].”
Most of the existing researches on thangka art are from the perspective of the development history of thangka art and culture and tend to focus on the research of thangka art itself and its macro history, especially the static study of culture, and lack of understanding of thangka art. Regarding the inheritance of thangka art, more attention is paid to the education of monasteries in Tibetan areas before liberation, while the inheritance of thangka art is constantly changing. However, thangka art is currently facing the challenge of cultural globalization, and the living space of thangka art inheritance is also facing the impact from foreign cultures. Therefore, the inheritance of thangka art should be considered from the perspective of promoting life and human development, so as to facilitate its protection and inheritance.
As a method of data processing and analysis, the essential difference between data mining and traditional data analysis (such as query, report, and online application analysis) is that data mining is to mine information and discover knowledge without explicit assumptions. The information obtained by data mining should have three characteristics: unknown in advance, effective, and practical [21]. The difference between it and the traditional statistical method is mainly reflected in the following: the usual statistical method is based on the existing assumptions and is verified from a large amount of data, while data mining is to obtain a new model from a large amount of data. Data mining is purely data-driven, while statistical methods are more about introducing human factors and analyzing them [22]. Exploratory data analysis is the branch of statistical methods most similar to data mining, but it is still oriented on data sets that are much smaller than data mining objects.
3. Methods
3.1. Problems with Raw Data
In the research process, if there is a lot of redundant information, it becomes more and more difficult to protect the knowledge of the original data during the crime. Data editing includes data correction, data integration, change, and function exit and selection. The purpose of data processing is to obtain the whole training course.
The original data is affected by the lack of parameters to a certain extent. In order to improve data quality, efficiency, and user friendliness, a series of operations need to be carried out in the original database, such as the preparation and transformation of the original data set.
3.2. Methods of Data Preprocessing
The data that needs to be analyzed by DM techniques may be incomplete data missing some attribute values or only containing summary data, noisy data with wrong attribute values, inconsistent data, and so on. These three are the three elements that measure data quality.
However, in practice, these data attributes are easily encountered. There are a number of factors that can lead to data incompleteness. For example, in market registration information, some data may not be included because it is not considered an important piece of information at the time of registration. For a number of reasons, the relevant data may not be recorded, and data inconsistent with other recorded data may have been deleted.
The data of some transactions may have inconsistencies, and some of the inconsistent data can be manually corrected through external references.
Metadata in a database refers to other data related to a dataset. Redundancy is another important issue, if one property can be derived from another then it may be redundant. Figure 1 presents a schematic diagram of data preprocessing.
For some tuples with missing attribute values, inference may be required to find missing values in the data. The data can also be noisy, with wrong attribute values. For example, the wrong data collection tool may be used, human error or computer error may occur in data entry, and errors in data transmission may occur from time to time. Technical limitations such as coordinating synchronous data transfers and occupying limited buffer sizes or using naming conventions that are inconsistent with data codes are all reasons for poor data quality. By perfecting the above three elements to clean the data, the low quality of the data will lead to the chaos of the DM process. Throughout many DM processes, there are procedures for handling “dirty data”, but they are not always robust. Instead, they might focus on avoiding matching the data to the function being modeled. Therefore, robust data cleaning routines are urgently needed.
3.3. Improvement Strategies for Data Preprocessing
The raw data is usually disorganized, and the data in the training set may contain data at various levels, with a wide range of data, and even differences in orders of magnitude. Since the degree of influence of features at different levels is different for the neural network, training on this kind of data will not only slow down the training speed, but may also cause the training results to be less accurate and less accurate. Therefore, before preprocessing the input data, it is necessary to perform some operations on the input data and normalize the training data, which makes it more convenient for data preprocessing and analysis of samples of different categories and improves the training speed and prediction model.
For those data with large differences in magnitude, numerical normalization can be performed on them, and the data after the operation will still be very different in magnitude. So, simply use numerical normalization to predict the data. The specific operation process diagram is shown in Figure 2.
3.4. Improved Batch BP Learning Algorithm
All supervised learning algorithms rely on error correction rules to improve the performance of the system. In the MLP model, the error obtained from the model is generally used to define the cost function of the network weight vector, and the network weight is used as its coordinate in space, and it is described as a multi-dimensional error performance surface. Over time, the lowest point of the error surface must be found continuously downwards to improve performance. To achieve this goal, the instantaneous gradient vector of the error surface needs to be calculated.
Batch learning is an offline deep learning method, and its process can be described as follows. Each time the network is trained with all input samples, for a training sample, we compare the error between the network output and the target output, and then find all the input samples [23–25]. After the error corresponding to the sample, the error is averaged to obtain the error value of the entire training sample, and then the gradient is obtained by using this error, and then the network parameters are corrected [26, 27].
The weight vector of the adjusted neural network will produce new points that return to the error surface. Part of Batman training is to adjust weight after training all the training data. Therefore, this error is defined as the average running error in all training equipment. That is, this experiment realizes its weight through continuous learning. By correctly calculating its gradient and adjusting the network weight, it is possible to compress the cost function as much as possible [28–30]. Therefore, using the batch learning process to accurately calculate the cost vector, the vector can be quickly converted to the minimization of the cost function. In this article, ten transformation techniques are used to adjust the training data, so that the other nine kinds of training can be trained and adjusted in the next step.
The batch learning BP algorithm designed and improved in this paper is mainly divided into two processing stages: feedforward stage and backpropagation stage. In the feedforward stage, the input samples are passed in from the input layer, processed layer by layer in the hidden layer, and then transmitted to the output layer. When the actual output value of the output layer does not match the desired result, it is corrected by back-propagating the error. The backpropagation of the error is to back-propagate the output error layer by layer to the input layer through the hidden layer, and then distribute the error to all neurons passing through each layer and obtain the error signal of each layer unit. Then, the obtained error signal is used to correct the weight of each unit. When the output error is within the controllable range, that is, the error is less than a certain set threshold, the cycle ends. In this paper, the momentum scalar factor is introduced, the weights are defined by the use of the momentum term factor, and the batch learning technique is used to make the obtained weights more accurate.
The training data X with P samples is given bywhere xp is the input vector of a sample p with n features (or dimensions)
And dp is a vector of desired outputs associated with it
Two differentiable activation functions are used in the algorithm: the hyperbolic tangent function and the logistic sigmoidal function. The output matrix y of the hyperbolic tangent activation function in the hidden layer is as follows:
It is worth noting that the zero vector is in the last column of matrix yt, which is the derivative of the fixed bias term. For a sigmoidal activation function, the equivalent equation is
In addition, when the output layer neurons are nonlinear, the batch BP algorithm must be used to iteratively learn the output layer weights W. Apply the hyperbolic tangent function in the output layer, a (P,K) dimensional matrix o is its output
To speed up the learning process, the momentum scalar factor ηm is introduced in this paper. The weights are updated through the use of momentum term factors, and, in order to be able to use them in the adaptation process of the network weights, the previous batches ▽V and ▽W are stored in ▽V_old and ▽W_old, respectively. Therefore, by using the LMS algorithm and the momentum term factor to adjust the network weights V and W, the rules are as follows:
4. Analysis of Results
4.1. Quantitative Analysis of Chinese Folk Painting Art and Spiritual Needs
With its unique art form, thangka transcends the boundaries of language and nationality. People can gain their own understanding of Buddhism and Tibetan Buddhism through the appreciation of thangka works. The special production process and coloring pigments of thangka will keep it well preserved for a long time, so it will not only benefit contemporary people and society, but will also make great contributions to mankind in the long-term future. In addition, different from the mural art, thangka is limited to a certain fixed place [31–33]. It is easy to carry and easy to move and hang. It can realize its traveling exhibition in the country and even around the world [34–36]. Coupled with the combination of picture albums, picture books, animations, and other forms in the later period, more people can experience the spiritual core that thangka makers want to convey, experience Chinese culture, and enter the world of Buddhism. Therefore, the breakthrough of thangka in language, time, and space will benefit more people.
The original data of a group of experiments did not undergo the operation of order-of-magnitude normalization, and the data preprocessing operation of normalization was directly performed. Another set of experiments is to perform order-of-magnitude normalization operations on the original data, followed by normalized data preprocessing operations. After unifying the order of magnitude of the data to a certain range, other routine operations on the data are preprocessed. After these preprocessing, the batch learning BP neural network algorithm to be mentioned later is applied to obtain the experimental results. The relationship between the contribution rate of Chinese folk painting art and spiritual needs is shown in Figure 3. The contribution rate of Chinese folk painting art and the accuracy of spiritual demand prediction are shown in Figure 4.
For the experimental group, which did not use order-of-magnitude normalization operations, the predicted value obtained by applying the BP neural network algorithm to learn has a large error with the actual value. For the experimental group that uses order-of-magnitude normalization operations, the error between the predicted value and the true value obtained by applying the BP neural network algorithm for learning is smaller. It can be seen that the order of magnitude normalization operation makes the results of data mining have higher accuracy and better learning performance.
The function of culture is that “it can satisfy people's needs” and “human beings can use it in their own activities.” People need food, clothing, shelter, and transportation to obtain their biological needs. The emergence of art culture lies in meeting the basic needs of people's strong feelings for sound, color, and shape, enhancing self-confidence in emotional experience, developing moral habits, and having optimistic confidence and attitude to actively cope with difficult problems. Therefore, improving mental support and psychological support for Chinese society and our nation is a manifestation of the role of culture. Different nations have created civilizations and formed cultures in the course of historical development. Culture plays an important role in the development of nations. When culture is compatible with the development of nations, culture builds up the common consciousness of nations and promotes the development of nations.
Psychology analyzes personality traits through four parts: attitude traits, rational traits, emotional traits, and volitional traits of traits. From a static point of view, a nation's character structure includes the attitudinal, intellectual, emotional and volitional features of the nation's character. From “Guru Padmasambhava's Thousand Thangkas”, we can also see that the attitude characteristics of the Tibetan character are mainly manifested in love of hometown, hard work, self-improvement, integrity, harmony with nature, and so on. The intellectual characteristics of the national character are the psychological characteristics of the nation in its cognitive activities.
If you want to deeply understand Tibetan culture, you cannot avoid Tibetan Buddhism. And to understand Tibetan Buddhism, you must understand the most important figure in the history of Tibetan Buddhism's development- Guru Rinpoche. In the preprosperous period of Tibetan Buddhism, Guru Rinpoche incorporated the Bon deities into the Buddhist Dharma Protector System, formally established the foundation for the spread of Buddhism in Tibet, and opened the prelude to the tantricization of Tibetan Buddhism. After a large number of eminent monks that promoted the esoteric method, many obstacles were cleared. Padmasambhava, as the founding leader of Tibetan Buddhism (tantric sect), plays a pivotal role in the development of Tibetan Buddhism and Tibetan culture. He has profoundly influenced the humanities, history, art, religion, customs, and other aspects of Tibet. However, for such a person who is so important to Buddhism and Tibetan culture, he is rarely known in the Han region and even in other parts of the world. This is a lack for the promotion of Buddhism and the spread of Tibetan culture.
At the same time, in modern society, with the improvement of human's ability to transform nature and create material, the pace of people's life is accelerating, and it is difficult for people to understand such a biography of Guru Rinpoche. Therefore, in order for the public to know and understand Guru Rinpoche, it must be done with media and methods that are adapted to the dissemination of knowledge in modern society and culture. In order to allow people to come into contact with and understand the legendary experience of Guru Rinpoche and to enable people to understand Tibetan Buddhism and Tibetan culture through understanding Guru Rinpoche's extraordinary Dharma-promoting career and to promote the excellent art of national painting, “Guru Padmasambhava's Thousand Thangkas” was born in response to demand. This artistic treasure will reproduce the karma of Padmasambhava, so that more people can deeply understand the culture and art of Tibet and the history of Tibetan Buddhism.
4.2. Quantitative Analysis of Chinese Folk Painting Art and Material Life
Thangka is a unique culture in Chinese Tibetan culture. Its content includes Tibetan politics, culture, history, and other aspects, so it is also called the encyclopedia of Chinese Tibetans. Therefore, it is very necessary to master the technique of thangka painting. When an artist draws a thangka, the first step is to choose a canvas of the right size and fit it into a small wooden frame according to the surrounding of the canvas. The color of the canvas should be mostly light, and the thickness and hardness should be moderate. Once the canvas is stabilized, it is time to prepare its “base color,” which is to apply a thin layer of glue to the canvas and take it out to dry. The next step is to draw the main positioning line, generally using gold to outline the outer layer of the picture and the edge of the rest of the picture, which is called “golden line.” The clothes and external patterns of these painted Buddha statues are mainly determined by these factors: which school the painter belongs to, his knowledge and proficiency in this school, the painter's painting ability, and so on.
Through the publication of “Guru Padmasambhava's Thousand Thangkas”, people will realize more clearly that Guru Rinpoche was the first to introduce tantra into Tibet, establish the harmony of tantra, and is the great founder and foundation of Tibetan Buddhism. Humans are fully enlightened Buddhas, the masters of Tantric Buddhism who appear in this world, the masters of all the Tathagatas, the incarnation of Amitabha, the master of the Western Paradise of Ultimate Bliss, the body of Amitabha, the words of Avalokitesvara, and the words of Sakyamu. The mind of Ni Buddha and all Buddhas is formed, and it is the manifestation of the body, speech, and mind of all Buddhas in the three times. This opens up people's admiration for Guru Rinpoche and takes this opportunity to study more deeply what is esoteric law, what is Tibetan Buddhism, and what is the relationship between Tibetan Buddhism and Chinese Buddhism. Figures 5 and 6 show the quantitative analysis results of Chinese folk painting art and material life.
4.3. Quantitative Analysis of Chinese Folk Painting Art and Individual Physical and Mental Development
With the development of high-tech in the electronic age, the way people in modern society hear and know information is undergoing tremendous changes, and its development has extended to the appreciation of works of art and the dissemination of cultural information. Everyone in today's society longs for the purest pacification of their hearts. However, most of the time, people are busy doing all kinds of obligations that they have to perform, fatigue their bodies, and waste their own lives. When drawing thangkas, the painters will relax their body and mind slowly, and their minds are full of sacred land and holy hearts. They keep drawing every day, so their hearts are also healthy every day. This is the painting “Guru Padmasambhava”. The prerequisites for Thousands of thangkas are also the sincere wishes of the painters.
Cognition refers to the acquisition of knowledge through mental activities. Cognitive process is the information processing process of individual cognitive activities, which is completed through cognitive activities such as individual perception, imagery, imagination, memory, and thinking. Creative thinking is an important part of cognitive activities, and it is a common way of thinking in “Guru Padmasambhava's Thousand-Painted Thangkas”. Painters must always grasp a clear way of thinking when drawing thangkas; and creative thinking is a must for painters. “Thousands of Thangkas by Guru Rinpoche” faithfully reflects the grandeur of Guru Rinpoche in those days through the art form of thangka. The process of spreading the Dharma shows the incredible legendary experience of the master. The painter will start thinking according to the life experience of the master when drawing, so as to vividly display the brilliant life of the master.
Using this data set, the improved batch learning BP algorithm and the traditional BP algorithm were trained and learned, respectively. The improved batch learning BP algorithm in this paper has higher classification accuracy than the traditional BP algorithm. For the data set in the experiment, the batch learning BP algorithm and the traditional BP algorithm are used for training and learning respectively, so as to obtain the prediction of the result of the classifier. The accuracy rate refers to the percentage of the ratio between the predicted value and the actual value (or the ratio of the ratio between the actual value and the predicted value, in order to facilitate the observation of the results, whichever is less than 1). By learning and predicting 11 data sets, the predicted algorithm accuracy is obtained respectively, and its average value is calculated. The average accuracy of the traditional BP algorithm is 91.80%, while the average accuracy of the batch learning BP algorithm is 93.16%. It can be seen that compared with the traditional BP algorithm, the classification accuracy of the batch learning BP algorithm is improved by 1.36%.
At the same time, due to the uneven quality of the data set, it directly leads to the difference in the accuracy of the two algorithms. Therefore, before data mining, the quality of the data must be strictly controlled to obtain a better classification effect. The correlation between the contribution rate of Chinese folk painting art and individual physical and mental development is shown in Figure 7. The contribution rate of Chinese folk painting art and the accuracy rate of individual physical and mental development prediction are shown in Figure 8. While the accuracy of batch learning BP neural network algorithm is improved, one of the disadvantages that follow is the increase in time consumption, which is also an inevitable problem, but a small increase in time cost is exchanged for accuracy. The degree of improvement, in a sense, is still desirable.
5. Conclusion
An improved method of data preprocessing is presented; that is, the data preprocessing part is improved before training starts. Before preprocessing, the order of magnitude of the data is normalized, so that the values of the data are basically in the same order of magnitude, so that a better data preprocessing effect can be obtained, and the effect and accuracy of the subsequent data mining process can be directly improved. Chinese folk painting art is not only the religious ideology of the Tibetan people, but also the carrier for the Tibetan people to communicate with heaven and earth. It contains cultural information of human aesthetic concepts, ways of thinking, religious beliefs, and customs in Tibetan areas. It has great historical value, religious value, and cultural information. Through the field investigation of the base of “Guru Padmasambhava's Thousand Thangkas,” this paper concludes that the main methods of inheritance of Chinese folk painting art are family and teacher inheritance. Along with the development of modern society, school inheritance and enterprise inheritance also appear. Through investigation and analysis, it is found that there are limitations and problems in the inheritance of Chinese folk painting art, such as insufficient knowledge of inheritors, the influence of social environment on the inheritance of Chinese folk painting art, the single inheritance content, and the crisis of inheritance methods. Finally, according to the problem, it is proposed that, in the process of inheritance of Chinese folk painting art, we should improve the understanding of the importance of inheritance of Chinese folk painting art, create a good inheritance atmosphere, enrich the content of Chinese folk painting art, and promote the diversification of Chinese folk painting art inheritance. The Chinese people, especially the young generation, should take the initiative to undertake the mission of inheriting China's precious intangible cultural heritage thangka and truly realize the diversified development of education.
Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by Pan Tianshou College of Architecture, Art and Design, Ningbo University.