Cultural Globalization: On the Reconstruction of Local Knowledge in Documentary Films on Non-Foreign Heritage in Henan, China

Intangible cultural heritage (ICH) is a valuable historical and cultural heritage of the Chinese nation, carrying rich cultural connotations and historical memories, and is of irreplaceable importance. Henan’s intangible cultural heritage is rich and diverse, and has had a significant impact on people’s ideology, moral concepts, and behavioral norms during China’s long agrarian era [1]. Henan folk literature is extremely rich, mainly myths, legends, stories, ballads, proverbs, etc., which can be called the epitome of Chinese folk literature. Henan folk operas have a long history and have always had an important position in the folk cultural life of Henan. There are also folk acrobatics, folk martial arts and folk music related to folk opera. In addition, there are also a large number of folk cultures that have survived in the folklore [2-3]. These broad and heavy national folk cultures are a rich cultural wealth left to us by our ancestors, and are important political resources, moral resources and cultural industry resources for the construction of socialist modernization nowadays.

Studying the current situation of safeguarding and utilization of intangible cultural heritage in Henan Province helps to gain a deeper understanding of the richness and value of intangible cultural heritage in the region, promotes the formulation and implementation of relevant safeguarding policies, and pushes the inheritance and development of intangible cultural heritage in the direction of more scientific and standardized work [4]. By analyzing the problems and challenges in the protection and utilization, effective countermeasure suggestions are put forward to provide reference and reference for the continuous promotion of the protection and utilization of intangible cultural heritage in Henan Province. Research on the current situation of the protection and utilization of intangible cultural heritage in Henan Province has important practical significance and far-reaching historical significance [5-6].

Literature [7] emphasized the importance of analyzing the spatial distribution characteristics and influencing factors of the intangible cultural heritage resources in the river basin of Henan Province for its protection, value enhancement and sustainable development. Using Tyson polygon, geographically weighted regression and other techniques, the study presents the distribution pattern of cultural heritage in Henan Province. And suggestions are made for its sustainable protection and utilization. Literature [8] analyzes Henan gongzi music and its dissemination so as to determine the authenticity of non-heritage preservation in the context of the digital era and to effectively understand the dissemination and preservation of traditional culture in contemporary society by emphasizing the important role of protecting and promoting cultural diversity. Literature [9] launched a systematic examination of the spatial distribution of 417 ICHO items and successors using methods such as the nearest neighbor index. The results of the study pointed out the spatial aggregation of ICHO and told about its distribution in various regions. Education and tourism were identified as potential initiatives for the preservation of ICHO, emphasizing the important role of integrating the inheritance in government, education, and other fields. Literature [10] outlines the current state of research on intangible cultural heritage tourism based on bibliometrics. It is pointed out that countries around the world in this field show a large gap, while China started late and fast, and should strengthen cooperation, expand research content, and improve research methodology in order to promote more scientific and considerable research. Literature [11] aims to explore the heritage resources in central China to help young people construct their identity and culture through patriotism education and ideological and moral education. For this region, the ideas and methods of application, inheritance and protection of cultural heritage need to be further researched in the light of the actual situation of the region. Literature [12] reveals that an important path to improve the management and protection of intangible cultural heritage is to study its spatial distribution characteristics and its influencing factors. Using standard deviation ellipse, spatial autocorrelation and other methods, the spatial and temporal distribution characteristics and causes were quantitatively analyzed with many national-level ICH in the Yellow River Basin as the research object and discussed. Literature [13] explored the advantages of higher education in the protection and inheritance of intangible cultural heritage, and proposed that local universities in Henan give full play to the advantages of relevant disciplines and specialties, in order to realize the nurturing role of colleges and universities and to promote the effective inheritance of the culture of the Central Plains. Literature [14] utilizes qualitative methods and cultural ecology theories aimed at examining historical and CC observations and conducting field surveys. A comparative test concluded that the openness of the social space is mainly reflected in the canal space as a linear cultural space of holistic interaction, and emphasized that the dance music of intangible cultural heritage in Henan Province is mainly percussion-based. Literature [15] takes Yuzu Shenhoujun porcelain intangible cultural heritage project as the research object, analyzes its history and culture based on literature review and other methods, and outlines its characteristics, production methods and patterns. It also puts forward the creative strategy of packaging design that combines the intangible cultural heritage with the local cuisine of Henan Province.

The study takes a certain non-heritage documentary presenting a local rare play in Henan as the object of discussion, and explores the creative performance of local knowledge in non-heritage documentaries in combination with its local knowledge content presentation. Subsequently, the online review data of several Henan non-heritage documentaries are collected, the topics are pre-extracted using LDA, the word vectors are obtained by Word2vec algorithm, and the high-frequency words of the text are clustered using the improved K-means algorithm, the topic clustering results of the online reviews are discussed, and the audience needs of the non-heritage documentaries are summarized. Then we select several Henan non-heritage documentaries as data samples, take the content theme, emotional connotation, character subject, background music, narrative technique and scene presentation as conditional variables, and take video likes, comments, retweets and favorites as outcome variables, and utilize fuzzy set qualitative comparative analysis method, and finally arrive at the realization of high dissemination effect through the analysis of necessity of single variable and the sufficiency analysis of multiple conditional grouping state by using fuzzy set qualitative comparative analysis method Grouping Path of Non-legacy Documentary.

2

Reconstruction of Local Knowledge in Documentary Films on Henan’s Nonheritage

“Local knowledge” is an important term in anthropological research, proposed by the famous anthropologist Clifford Geertz, emphasizing that local knowledge is a kind of knowledge with local status and inherent characteristics, that is, a knowledge system formed by local culture, which is produced in a specific geographical spatial unit and is closely related to the local race, ethnicity and community. In the face of the erosion of local cultures by cultural globalization, local knowledge systems provide an internal perspective that dissolves foreign cultures, allowing a region’s culture to maintain the stability of its core.

According to the creation characteristics and expression advantages of documentaries, as well as the theoretical content of local knowledge, a documentary about Huai Bang, a rare local drama in Qinyang, Henan Province, is analyzed as an example. The reconstruction of local knowledge in non-legacy Henan documentaries is shown in Figure 1. The focus is on the methods and forms used to express local knowledge in documentaries from the following aspects.

2.1

Identification and articulation of non-exhaustive cultural heritage

Huaiqing Opera is a type of vocal theater formed by the original Haishen Opera combined with the dialect of the Huaiqing Province area. Therefore, the documentary starts from the geographical environment of Qinyang and combines historical and humanistic narratives. The movie shows three representative natural cultural landmarks in Qinyang, namely the water of the Qinhe River, Shennong Mountain, and the Three Sacred Pagodas. The film cuts into the story from a microscopic point of view, searching for the intrinsic connection between Huayang artists, Huayang troupes and Huayang culture, and linking them together to re-explain the memories of the county-level non-heritage theater troupes.

2.2

Diverse spaces enrich narratives

The diverse narrative space can not only promote the development of the story, but also enrich the story content, which provides a multifaceted supplement for the presentation of the whole picture of intangible cultural heritage in the documentary of non-heritage. The documentary mainly shows the status quo of local traditional theater, the characteristics of opera performance and rehearsal, and the personal life of Huaidang singers, which makes the narrative space of the film present a multi-site and multi-dimensional feature. When focusing on the local opera, it is necessary to take the troupe’s workplace and the performance stage as the main space, and when filming in the troupe’s workplace, it is also subdivided into the troupe’s rehearsal hall, the troupe’s office, the troupe’s conference room, and other narrative scenes. The performance stage is also divided into two parts: the formal performance stage and the backstage.

2.3

Detailed interpretation of local knowledge

In the narrative process of non-heritage documentaries, it is very important to grasp the details and the smallest component unit in the work to depict the character, natural scenery, event development, social environment, etc. is the details. Creators need to grasp the emotions in the details of the narrative, penetrate the characterization from the details of life, and thus promote the development of the story. There are many details of the subject’s life captured in the documentary, such as in the shooting of the first rehearsal of a new play, the creators capture the dialogues and micro-expressions of the director and the actors, which makes the images and characters more realistic.

2.4

Civilianization perspective

The narrative strategy of a civilian perspective is a trend in the creation of non-heritage documentaries. Individual narratives from a civilian perspective are embedded in the stories of the groups that have gained their identities from them, thus expanding the space for viewers to construct their cultural memories. In the documentary, the daily life of the storytellers as ordinary members of the troupe is shown through the living narrative and civilianized perspective, presenting a natural sense of life, such as having a working meal with the troupe members in the interval of finishing the rehearsal work.

2.5

Expression of synesthesia

Simultaneous sound refers to the objectively existing sound recorded on the spot in the process of image shooting, which has strong authenticity and assumes the function of transmitting information and assisting the narrative of the picture in the non-heritage documentary. In the documentary, the plot of the troupe rehearsing a new play also uses a lot of simultaneous sound. In the traditional classic Huai Bang repertoire, singers are not allowed to change or innovate the classic songs, so the classic Huai Bang repertoire retains the localized classic tunes. However, with the innovative development of Huayang, the artists of Huayang usually add their own special singing style on top of the original style of Huayang, so in the rehearsal clips of the new plays, you can hear a more lively Huayang singing style.

3

Analysis of Henan’s non-legacy documentary audience needs

Intangible cultural heritage is the crystallization of the fusion of culture and art nurtured by the Chinese nation in the long history of culture, and the inheritance and development of intangible cultural heritage is an important measure for the continuation of historical lineage and the consolidation of cultural confidence. In recent years, the rapid development of the new media era and the increasing national attention to traditional culture, less appear in the public field of vision of the intangible cultural heritage subject documentaries gradually appeared in the public field of vision, non-heritage image archive for the transmission of the spirit of non-heritage and future research on non-heritage subject provides a valuable reference material. This chapter analyzes the audience demand of non-heritage documentaries in Henan, and obtains the audience’s demand for non-heritage documentaries through the theme extraction of online commentaries on documentaries, which provides reference for the production and dissemination of non-heritage documentaries.

3.1

Theme Extraction Methodology

3.1.1

LDA topic pre-extraction

The LDA model is proposed based on the Bayesian idea that the distribution of unknown parameters is not fixed and the sample distribution is determined. The basic process of Bayesian parameter estimation is: prior distribution + knowledge of the data = posterior distribution. And the multinomial distribution and the Delikere distribution are conjugate, therefore, the Delikere prior + multinomial distribution of the data’s → posterior distribution is the Delikere distribution. It can be expressed by the formula: (1) $D i r (\vec{p} | \vec{α}) + M u l t C o u n t (\vec{n}) = D i r (\vec{p} | \vec{α} + \vec{n})$

The frequency $\vec{n}$ of occurrence of each word obeys a multinomial distribution $M u l t (\vec{n} | \vec{p}, N)$ given the prior distribution $D i r (\vec{p} | \vec{α})$ of parameter $\vec{p}$ , the posterior distribution is: (2) $p (\vec{p} | W, \vec{α}) = D i r (\vec{p} | \vec{α} + \vec{n})$

After the posterior distributions of the parameters are available, the mean values are chosen as the estimates of the parameters in the parameter estimation. The probability density expression for the k-dimensional Dirichlet prior distribution of the multinomial distribution $\vec{p} = (p_{1}, p_{2}, \dots, p_{k})$ is: (3) $\begin{array}{l} D i r i c h l e t (p_{1}, p_{2}, \dots, p_{k} | α_{1}, α_{2} \dots, α_{k}) \\ = \frac{Γ (\sum_{i = 1}^{k} α_{i})}{Π_{i = 1}^{k} Γ (α_{i})} Π_{i = 1}^{k} p_{i}^{α_{k - 1}} \end{array}$

Further the output probability of text corpus W can be calculated as deduced in the following equation: (4) $\begin{matrix} p (W | \vec{α}) = \int p (W | \vec{p}) p (\vec{p} | \vec{α}) d \vec{p} \\ = \int \prod_{k = 1}^{v} p_{k}^{n_{k}} D i r (\vec{p} | \vec{α}) d \vec{p} \\ = \int \prod_{k = 1}^{v} p_{k}^{n_{k}} \frac{1}{Δ (\vec{α})} \prod_{k = 1}^{v} p_{k}^{α_{k + 1} - 1} d \vec{p} \\ = \frac{1}{Δ (\vec{α})} \int \prod_{k = 1}^{r} p_{k}^{n_{k} + α_{k} - 1} d \vec{p} \\ = \frac{Δ (\vec{n} + \vec{α})}{Δ (\vec{α})} \end{matrix}$

In the LDA model, K topic corresponds to K independent Dirichlet-Multinomial covariates. In the process of sampling out the topic distribution $\ddot{θ}$ by a priori parameter $\vec{α}$ and then generating the topic number z_m,n of the n th word in the document by $\vec{θ}$ , $\vec{α} \to \vec{θ}$ corresponds to the Dirichlet distribution and $\vec{θ} \to {\vec{z}}_{m}$ to the multinomial distribution, which can be obtained based on Eq. (5): (5) $p ({\vec{z}}_{m} | \vec{α}) = \frac{Δ (\vec{n_{m}} + \vec{α})}{Δ (\vec{α})}$

In the corpus, the topic generation process of M document is independent of each other, then the topic generation probability of the whole corpus can be introduced as: (6) $\begin{matrix} p (\vec{z} | \vec{α}) = \prod_{m = 1}^{M} p (\vec{z_{m}} | \vec{α}) \\ = \prod_{m = 1}^{M} \frac{Δ (\vec{n_{m}} + \vec{α})}{Δ (\vec{α})} \end{matrix}$

The derivation of the probability distribution for generating word w_m,n in K subject word distribution ${\vec{φ}}_{k}$ selected at subject number k = z_m,n is similar to the derivation process of the subject probability distribution described above, then the probability of word generation in the whole corpus can be deduced as: (7) $p (\vec{w} | \vec{z}, \vec{β}) = \prod_{k = 1}^{K} \frac{Δ (\vec{n_{k}} + \vec{β})}{Δ (\vec{β})}$

In summary, the joint probability distribution of the subject words can be obtained: (8) $p (\vec{w}, \vec{z} | \vec{α}, \vec{β}) = \prod_{k = 1}^{K} \frac{Δ (\vec{n_{k}} + \vec{β})}{Δ (\vec{β})} \prod_{m = 1}^{M} \frac{Δ (\vec{n_{m}} + \vec{α})}{Δ (\vec{α})}$

For a given set of documents, w_m,n is an observable value, a and β are empirically given a priori parameters, and the other variables θ_m , z_m,n , and ϕ_k are unknown implicit variables, which need to be learned and estimated from the observed data.

3.1.2

Word2vec word vector acquisition

LDA topic model for the semantic representation of the document is a word probability distribution, ignoring the deep semantic information, in order to better improve the topic representation, the topic words can be converted to vector representation, and then be able to characterize the text topic into vector form for further quantitative calculations.

The training structure of Word2vec model: CBOW model which predicts the current word according to the contextual lexical items and Skip-gram model which predicts the contextual lexical items according to the current word. The learning objectives of the two models are represented as follows: (9) $L_{C B O W} = \sum_{w \in C} \log p (w | C o n t e x t (w))$ (10) $L_{S t i p - g r u m} = \sum_{w \in C} \log p (C o n t e x t (w) | w)$

where w is the center word, Context(w) is the contextual lexical item of the center word, and C is the corpus.

As an example, the CBOW model using Huffman tree optimization solves the problem of defining Eq. (9) by projecting layer vectors and Huffman trees. Eq. (11) is the j th binary classification probability of the Huffman tree: (11) $p (d_{j}^{w} | X_{w}, θ_{j - 1}^{w}) = {[σ (X_{w}^{T} θ_{j - 1}^{w})]}^{1 - d_{j}^{w}} \cdot {[1 - σ (X_{w}^{T} θ_{j - 1}^{w})]}^{d_{j}^{w}}$

where $d_{j}^{w}$ denotes the Huffman take of the j nd node of the target word, X_w is the mean vector of the context word, $θ_{j - 1}^{w}$ is the model parameter of the j – 1 th node of the target word, and l^w is the maximum number of nodes corresponding to the Huffman encoding of the target word. Then: (12) $\begin{matrix} L_{C B O W} = \sum_{w \in C} \log \prod_{j = 2}^{J^{m}} {{[σ (X_{w}^{T} θ_{j - 1}^{w})]}^{1 - d_{j}^{*}} \cdot {[1 - σ (X_{w}^{T} θ_{j - 1}^{w})]}^{d^{ψ}}} \\ = \sum_{w \in r} \sum_{j = 2}^{r^{κ}} {(1 - d_{j}^{w}) \cdot \log [σ (X_{w}^{T} θ_{j - 1}^{w})] \\ + d_{j}^{w} \cdot \log [1 - σ (X_{w}^{T} θ_{j - 1}^{w})]} \end{matrix}$

C is the number of context words, the corresponding v_w is the input vector, W is the weight matrix, and v_w is a row of W. The maximum value of the objective function for solving Eq. (12) is obtained by making the objective function partialized with respect to θ: (13) $\begin{matrix} \frac{\partial L (w, j)}{\partial θ_{j - 1}^{w}} = \frac{\partial}{\partial θ_{j - 1}^{w}} {(1 - d_{j}^{w}) \cdot \log [σ (X_{w}^{T} θ_{j - 1}^{w})] \\ + d_{j}^{w} \cdot \log [1 - σ (X_{w}^{T} θ_{j - 1}^{w})]} \end{matrix}$

The sigmoid function used for the binary classification of the Huffman tree, for the sigmoid function has the characteristics of equation (14): (14) $σ' (x) = σ (x) [1 - σ (x)]$

Bring Eq. (14) into Eq. (13) and combine and simplify to get Eq. (15): (15) $\frac{\partial L (w, j)}{\partial θ_{j - 1}^{w}} = [1 - d_{j}^{w} - σ (X_{w}^{T} θ_{j - 1}^{w})] X_{w}$

Similarly, the partial derivative of the objective function with respect to X_w can be obtained as: (16) $\frac{\partial L (w, j)}{\partial X_{w}} = [1 - d_{j}^{w} - σ (X_{w}^{T} θ_{j - 1}^{w})] θ_{j - 1}^{w}$

The CBOW model learning process uses a full batch stochastic gradient ascent method, which can be obtained according to the knowledge related to gradient ascent, and the update for $θ_{j - 1}^{w}$ is shown in Eq. (17): (17) $\begin{matrix} θ_{j - 1}^{w} : = θ_{j - 1}^{w} + η \frac{\partial L (w, j)}{\partial θ_{j - 1}^{w}} \\ = θ_{j - 1}^{w} + η [1 - d_{j}^{w} - σ (X_{w}^{T} θ_{j - 1}^{w})] X_{w} \end{matrix}$

In word2vec model, X_w is the sum of word vectors of context words, not the word vectors of individual words, for the update of word vectors of individual words, the updating process is to directly apply all the updates of X_w to the word vectors of each word, then the update of v(w) is as shown in equation (18): (18) $\begin{matrix} v (w) : = v (w) + η \sum_{j = 2}^{r} \frac{\partial L (w, j)}{\partial X_{w}} = v (w) \\ + η [1 - d_{j}^{w} - σ (X_{w}^{T} θ_{j - 1}^{w})] θ_{j - 1}^{w} \end{matrix}$

η is the learning rate, the process of predicting contextual word information from the center word is achieved by the above derived CBOW model. The word vectors obtained after word2vec model training can represent the similarity relationship between words through the distance relationship on the semantic space.

3.1.3

Improved K-means algorithm

In this section, from the perspective of semantic similarity of topic words and elimination of topic ambiguity, we propose a method that proposes a topic representation method based on Word2vec model and LDA model as a way to find the initial clustering center vectors of K-means. Can effectively utilize the topic information extracted from the LDA model and word vectors Word2vec model word to word similarity semantic information and semantic order information. Based on the LDA model and Word2vec model weighted and improved characterization of the subject vectors of the algorithm flow is shown in Figure 2. The specific design steps are as follows: 1)

Set the number of potential topics k , hyperparameters a and P for the text data source.

2)

Use the preprocessed corpus data as the input to the LDA model, train the model, and get the topic lexical item information.

3)

Select the first n words of each topic T_i , as the topic words, and calculate the contribution of the topic words to the topic, i.e., calculate the weight of the word w to the topic.

4)

Map the topic words to the semantic space by Word2vec model training results.

5)

The topic vector is expressed as the weighted sum of word vectors within a topic and the weights of words occupying the topic, i.e., for a topic, the word vector of each topic word within the topic is multiplied by the weight of the word occupying the topic.

For long text data sources, the LDA model is used as the input for topic clustering after obtaining the topic lexical information and combining it with the Word2vec model for topic expression.

3.2

Comment capture

Ten Henan non-legacy documentaries are selected as the research object, and the text of user comments on multiple video platforms is collected as the data source, which includes the user name, comment content, and comment time, resulting in a total of about 4,000 popular comments with a total word count of nearly 80,000 words. In the data processing stage, duplicate comments are firstly removed, followed by calling Python’s Jieba Segmentation Package and LTP Natural Language Processing to segment the comment content, and then introducing the deactivation word list and adding deactivation words appropriately according to the study.

3.3

Thematic clustering results

The data is incorporated into documents and topic clustering is performed using the topic extraction model proposed in this paper. Because of the need to autonomously define the number of topics, this paper uses the confusion degree to indicate the optimal number of topics. Confusion degree indicates the uncertainty of the topic to which the document belongs, in general, when the confusion degree is the lowest, it is the optimal number of topics, but it does not mean that the more the number of topics is the better, the number of topics is too much will be overfitting the model phenomenon. The variation of the confusion degree and the number of themes is shown in Figure 3. When the number of topics is equal to 4, it is the “elbow-shaped” low point, the difference between it and the corresponding confusion degree when the number of topics is 3 is large, and the difference between it and the confusion degree when the number of topics is 5 is very small. When the number of themes is 8, the confusion reaches its lowest value. Therefore, both 4 and 8 can be identified as the optimal number of topics.

Accordingly, this paper assumes that 4 and 8 are the optimal number of topics, uses the model for topic clustering, and compares and analyzes the visualization results with the PyLDAvis package to decide the optimal number of topics. In the presented visualization results, the circle represents the theme, the size of the circle indicates the number of documents corresponding to the theme, and the smaller the distance between the circles, the higher the similarity between the themes, and the worse the theme clustering effect.

When the number of topics is 4, the visualization results are shown in Fig. 4, Fig. (a) shows the distribution of each topic, and Fig. (b) shows the 20 most relevant lexical items for topic 1. The frequency of the 20 items in Theme 1 is 0.1%~9.5%, and the proportion of “culture”, “spirituality” and “inheritance” in the review of intangible cultural heritage documentaries is the highest, all of which are above 7.5%. Different themes are farther apart and there is no overlapping of themes. When there are 8 themes, the phenomenon of multiple theme stacking occurs. After thorough analysis, this paper selects 4 as the optimal number of themes for this study.

3.4

Analysis of audience needs

According to the clustering results, the four themes and their related characteristic words are shown in Table 1. The four themes results of clustering are summarized in terms of their thematic content, and it is found that certain needs of the audience are satisfied based on the content, characters, photographic production and other elements of the Henan non-heritage documentaries. This echoes with the theory of use and satisfaction that the audience contacts the media based on certain needs, and the audience’s needs have a constraining effect on the communication effect, and in order to obtain the expected communication effect, it is necessary to understand and satisfy the audience’s needs as much as possible, so this paper summarizes the content of the four themes based on the theory of use and satisfaction.

Table 1.

Four themes and their relevant feature words

Topic	Audience demand	Feature words	Proportion
Topic 1	Emotional resonance	Culture, spirit, China, professional, thing, grateful, respectable	31.92%
Topic 2	Knowledge acquisition	History, process, video, craftsmanship, inherit, knowledge	26.61%
Topic 3	Self-realization	Majors, documentary, learning, lifetime, hope, excellent	16.12%
Topic 4	Aesthetic acceptance	Director, shooting, aside, picture, language, documentary	25.35%

Combining the results of theme clustering and research needs, four demand dimensions were identified: emotional resonance demand, knowledge acquisition demand, self-realization demand, and aesthetic acceptance demand. Emotional resonance refers to emotions such as moving, crying, heartache, regret, and so on. that the audience generates during the process of watching the documentary, and it also includes the audience’s sharing of values about the current situation of Henan’s Nongenetic inheritance and protection strategies based on the content of the documentary and their own experiences, which corresponds to Topic 1. Knowledge acquisition demand refers to the cultural knowledge that the audience obtains from the viewing of Nongenetic documentary, which includes the cultural knowledge that the documentary directly expresses, and also the cultural knowledge that the audience obtains from social interactions. The demand for self-realization refers to the audience’s willingness to contribute to the protection of NHM through their own abilities or to engage in related professions after watching NHM documentaries, corresponding to Topic 3. The demand for aesthetic acceptance refers to the audience’s demand for the aesthetics of NHM documentaries in terms of cinematography, narration, images, etc., and whether or not it is consistent with the cognition, corresponding to Topic 4. Aesthetic acceptance demand refers to the audience’s demand for the aesthetics shown in the photography, narration and images of NRM documentaries, and the demand for their compatibility with cognition, corresponding to Topic 4. The audience’s demand for emotional resonance in the NRM documentaries in Henan is the largest, accounting for 31.92%, followed by the demand for knowledge acquisition and the demand for aesthetic acceptance, which account for 26.61% and 25.35%, respectively, and the need for self-actualization is relatively small.

4

Analysis of the factors affecting the dissemination effect of non-heritage documentaries

In order to better inherit the intangible cultural heritage of Henan, on the basis of exploring the audience demand of Henan’s non-heritage documentaries, the factors influencing the dissemination effect of its non-heritage documentaries are analyzed. Based on the number of broadcasts, likes, comments and favorites, the cases of Henan’s non-heritage documentaries are collected on relevant video website platforms, and the factors influencing the dissemination effect of non-heritage documentaries are analyzed after the observation of their content elements.

4.1

Research variables

4.1.1

Conditional variable design

There are many factors that affect the dissemination effect of Henan’s non-heritage documentaries, both internal factors of the documentaries and external factors such as science, technology, and policies. Considering that the number of condition variables in the QCA method should be appropriate for the number of samples and the availability of research data. This study chooses the internal factors of non-heritage documentaries for research, i.e., to study the influencing factors of the dissemination effect of Henan non-heritage documentaries from the quality of the documentaries themselves.

The finalized number of cases of Henan NRL documentaries in this dissertation is 35, and the conditional variables are selected as the content theme, emotional connotation, character subject, background music, narrative technique and scene presentation of the documentary, totaling six conditional variables.

4.1.2

Outcome variable design

The data that can visualize and quantify the dissemination effect in documentaries include the number of likes, comments, favorites and shares, and these four broadcasting data of the case samples of Henan non-heritage documentaries belong to the public data, which further ensures the transparency of this study. Therefore, these four data are used to calculate the dissemination effect of Henan non-heritage documentaries and as outcome variables to examine the factors influencing the dissemination effect of Henan non-heritage documentaries. The dissemination effect of each documentary is obtained by calculating the four indicators of likes, comments, favorites, and retweets of 35 Henan non-heritage documentaries according to the steps of the entropy weight method.

4.2

Research methodology

4.2.1

Methodological choices

Relying on the foundation of Boolean algebra, the qualitative comparative analysis method focuses on analyzing correlations between members of different sets in order to identify common properties across multiple cases. In other words, this analytical method explores general associations between sets by comparing numerous cases and applying Boolean logic to derive causal relationships. The Boolean method is a method of establishing Boolean logic truth tables based on dichotomy and set theory, and studying the occurrence of results in combination with sufficiently necessary conditions.

In this paper, the fuzzy set qualitative comparative analysis (fsQCA) is used to analyze the influencing factors of the communication effect of Henan non-heritage documentaries. First, this method integrates the advantages of quantitative and qualitative analysis, focusing on the multiplicity and complexity of causal relationships. Second, this method effectively sort out the logical associations among variables and identify parallel causal action mechanisms by examining multiple possible paths to the same outcome, thus analyzing multivariate interactions in a complex context.

4.2.2

fsQCA methodology

fsQCA overcomes the limitations of traditional csQCA in dealing with variables (only being able to analyze the causality of binary variables), and is not limited to dealing with many types of variables including continuous variables, and is a further extension of the range of applications based on csQCA. fsQCA introduces the metric of the degree of affiliation, which allows the researcher to numerically characterize the intensity of a particular state. Larger values of the degree of affiliation indicate a higher probability of occurrence of that state. Smaller values imply a lower probability of occurrence of the state. In general, the numerical value of the degree of affiliation ranges between 0 and 1, where 0 indicates no affiliation, 0.5 represents a neutral or indeterminate affiliation state, and 1 represents full affiliation. The affiliation degree was then used to construct a truth table, focusing on the values of consistency and coverage. The numerical analysis of consistency and coverage determines the form in which each conditional construct exists. Consistency serves as a key indicator for assessing whether a conditional variable satisfies sufficient or necessary conditions, while coverage reflects the comprehensiveness of the conditional constructs in explaining the results. The formulas for consistency and coverage are, respectively: (19) $C o n s i s t e n c y (X i \leq Y i) = \frac{\sum (\min (X i, Y i))}{\sum (X i)}$ (20) $C o n v e r a g e (X i \leq Y i) = \frac{\sum (\min (X i, Y i))}{\sum (X i)}$

where Xi denotes the degree of affiliation of case i to antecedent condition X . Similarly, Yi denotes the degree of affiliation of case i corresponding to outcome variable Y .

4.2.3

Data calibration

In FSQCA, data calibration is the process of assigning pooled affiliation scores to cases, so that each condition and outcome of the case has an affiliation score. The quantile value method was used to determine the three anchor points, using the 95%, 50% and 5% quantile values as the calibration anchors for each variable, and Ragin’s direct method was used to complete the calibration of the data, after calibration, all the variables were continuous values ranging from 0 to 1. Due to the existence of sample intersections with values calibrated to exactly 0.5, this paper adjusts 0.5 to 0.499 or 0.501 depending on the bias belonging to the intersection value, and this paper chooses to adjust the value that is at the intersection point to 0.501.

4.3

Necessity analysis

After assigning and calibrating the case samples of Henan Cultural Heritage documentary according to the coding setting standard at the level of video content elements, “Necessary Condition” was selected in the fsQCA software for calculation, and the results of the single-factor necessity analysis of high propagation effect are shown in Figure 5, and “~” means “non”, that is, “non-existent”.

At the level of the elements of non-heritage documentaries, the consistency level of the presence and absence status of each conditional variable is less than 0.9, so the absence of a single conditional variable is a necessary condition for the occurrence of the results. This result shows that content theme, emotional connotation, character subject, background music, narrative technique and scene presentation are not necessary conditions that affect the high communication effect of Henan’s non-heritage documentaries, and they cannot explain the reasons for the formation of the results individually. In the necessity analysis of individual condition variables, the consistency score of “emotional connotation” is 0.829, which is higher than 0.8 and less than 0.9, and the coverage rate is more than 0.5, which can be regarded as a sufficient condition. As far as this study is concerned, the existence of behaviors such as rich emotional elements and strong emotional expression implies the existence of “emotional connotation”, and the existence of this conditional variable has a certain role in promoting the dissemination of Henan’s non-legacy documentaries, which plays a key role in the occurrence of the results of high dissemination effect.

After analyzing the results of univariate necessity analysis of high dissemination effect, the univariate necessity analysis of non-high dissemination effect is carried out, and the results of univariate necessity analysis of non-high dissemination effect are shown in Fig. 6. The consistency level of the existence and non-existence status of each conditional variable is less than 0.9, which indicates that each conditional variable is not a necessary condition affecting the non-high dissemination effect of the Henan non-heritage documentaries, and the results can not be individually explained by the reasons. Reasons. In the necessity analysis of individual conditional variables, the consistency score of “character subject” is 0.823, and the coverage rate is 0.750, which can be regarded as a sufficient condition, indicating that the non-high dissemination effect of Henan non-heritage documentaries is greatly influenced by the conditional variable of character subject with strong professionalism, which means that in more than half of the cases, although the non-heritage documentaries have a high dissemination effect, they are not necessary conditions to explain the reasons for the formation of the results. cases, although the character subjects of the non-heritage documentaries have professional backgrounds, which should be an advantageous factor for the dissemination of Henan non-heritage documentaries, it leads to the documentaries not achieving better dissemination effects.

4.4

Sufficiency analysis

After the necessity analysis of individual conditional variables, this section carries out the sufficiency analysis of conditional grouping, i.e., the multifactor grouping path analysis, which is the key issue that QCA research can address. In this study, the given outcome refers to the realization of a high dissemination effect of non-heritage documentaries from Henan, and the antecedent elements refer to individual condition variables. The necessity analysis of single conditional variables is to explore whether there exists a certain antecedent factor that will definitely make Henan non-heritage documentaries realize a good dissemination effect, and the sufficiency analysis of conditional grouping is to explore which groupings of antecedent factors can make Henan non-heritage documentaries realize a good dissemination effect.

The grouping results of Henan’s non-legacy documentaries are presented as shown in Table 2, with “•” indicating that the core condition exists, “◊” indicating that the core condition is missing, “○” indicating that the marginal condition exists, and “⊗” indicates that the marginal condition is missing, and “blank” indicates that the presence or absence of the condition is irrelevant. According to the results of the analysis, there are 6 groupings of high dissemination effect in Henan non-heritage documentaries, and the consistency of the 6 paths is 0.825, 0.816, 0.821, 0.841, 0.827, 0.813, respectively, which is greater than the minimum standard of 0.75, indicating that the 6 groupings are a sufficient condition for high dissemination effect. The overall consistency of the 6 groupings is 0.927, which further illustrates that the 6 groupings are sufficient conditions for Henan’s non-legacy documentaries to produce high communication effects. The overall coverage of 0.758 for the six histograms is well above the acceptable standard of 0.3 and provides a strong explanation of the results.

Table 2.

The configuration results of the documentary series of Henan

Conditions	R1	R2	R3	R4	R5	R6
Subject content	○	○	•	•		•
Emotional connotation	⊗	○	○	•	•	○
Character subject	○	•	•		⊗	○
Back ground music	⊗	◊		○	⊗	◊
Narrative technique	○	•	○		○	•
Scene presentation	○	•	⊗	⊗	•
Raw coverage	0.187	0.186	0.127	0.068	0.096	0.166
Unique coverage	0.869	0.864	0.885	0.732	0.782	0.691
Consistency	0.825	0.816	0.821	0.841	0.827	0.813
Solution coverage	0.758
Solution consistency	0.927

5

Conclusion

Intangible cultural heritage is an indispensable part of the excellent traditional culture of the Chinese nation, and the intangible cultural heritage of Henan, China, is an important part of the local knowledge system. In this paper, under the background of cultural globalization, the concept of local knowledge is applied to the understanding of Henan intangible cultural heritage documentaries, and the intangible cultural heritage documentary about a certain Henan Huaidang opera genre is taken as an example to reconstruct and interpret, summarizing the creative expression of cultural identity, multi-space narrative, detail expression, civilian perspective and contemporaneous voice expression in multiple dimensions. And based on the inheritance and development of Henan’s non-heritage, the audience demand and communication influencing factors of its non-heritage documentaries are explored. On the one hand, a theme extraction method combining LDA theme algorithm, Word2vec algorithm and K-means algorithm is used to mine the online comments of the non-heritage documentaries, and four types of audience needs are summarized, namely, emotional resonance, knowledge acquisition, self-actualization and aesthetic acceptance, of which emotional resonance and knowledge acquisition take up the largest proportion of the needs, which account for 31.92% and 26.61%, respectively. On the other hand, the fuzzy set qualitative comparative analysis method is used to analyze the collected cases of Henan non-heritage documentaries, and six grouped paths are obtained, with the overall consistency and overall coverage of 0.927 and 0.758 respectively, which can explain most of the cases of high communication effect of Henan non-heritage documentaries, and the relevant production and publicity personnel can refer to the results of the grouped paths, focusing on the content theme of the non-heritage documentaries, the emotional content, character body, narrative technique and aesthetic acceptance. Relevant producers and publicizers can refer to the results of this group path, focusing on the content theme, emotional connotation, character subject, narrative technique and scene presentation of the non-heritage documentaries to enhance the communication effect of the non-heritage documentaries in Henan, and further promote the development of intangible cultural heritage in Henan.

Idioma:: Inglés

Calendario de la edición:: 1 veces al año
Temas de la revista:: Ciencias de la vida, Ciencias de la vida, otros, Matemáticas, Matemáticas aplicadas, Matemáticas generales, Física, Física, otros

RSS Feed de revista

Cultural Globalization: On the Reconstruction of Local Knowledge in Documentary Films on Non-Foreign Heritage in Henan, China

Hexi Wang

Mingjie Wang

Publicado en línea: 21 mar 2025

Recibido: 24 oct 2024

Aceptado: 04 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0637

Palabras claveLDA, Word2vec, K-means, fsQCA, Local knowledge, Henan intangible heritage documentaries

© 2025 Hexi Wang et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Palabras clave
LDA, Word2vec, K-means, fsQCA, Local knowledge, Henan intangible heritage documentaries