Pattern Identification and Behavioral Analysis of Sports Marketing Strategy Implementation in Big Data Environment

Sports marketing refers to a marketing approach to promote brands, products or services through sports events and activities. With the booming development of the sports industry, more and more enterprises begin to pay attention to sports marketing and take it as an important means to enhance brand image and expand market share [1-4]. Sports marketing is not an overnight process, enterprises need to establish long-term cooperative relationships with sports programs and events, and continuously invest resources in order to achieve good marketing results [5-7]. At the same time, enterprises can also work with other brands, media and other partners to carry out sports marketing activities, to achieve resource sharing and improve marketing effectiveness [8-10].

Sports marketing strategies include sponsorship and cooperation, brand activities and sports promotion, social media and content marketing, stadium advertising and event broadcasting, and sports star endorsement and brand image building. Brands need to be closely integrated with sports to enhance their visibility through sponsorship of events and athletes [11-14]. When formulating marketing strategies, brands need to choose their own sports marketing strategies according to their own characteristics and the needs of their target audiences, and make constant adjustments and optimizations based on market feedback [15-16]. Interact with users through interactive marketing and new media platforms such as social media to enhance participation and stickiness. Build brand image and reputation through differentiated competition and social responsibility. Attract users’ attention and increase exposure by producing high-quality sports content [17-20]. Sports marketing strategy is important for enterprises to stand out in the highly competitive sports industry, only through scientific and effective sports marketing strategy, brands can stand out in the fierce market competition and achieve greater commercial success [21-24].

This paper uses sports consumers’ purchase reviews for pattern recognition and behavioral analysis. Being in the position of fuzzy attributes to daily language, the method of fuzzification calculation is used to calculate the emotional intensity of the review words, and the emotional tendency is classified into positive and negative with a total of 9 levels. Extreme, high, and medium-degree words are regarded as centralized operators, while low and trace-degree words are regarded as weak operators, and the affiliation function of the word is constructed.The method of lexical inversion and then halving is used to defuzzify words modified by negative words, to measure the degree of sentiment shift, and complete the inference according to fuzzy inference rules. By defining online reviews as linguistic variables, a fuzzy semantic model is constructed, and the purchase behavior model of sports goods is established according to the emotional semantic resources in the reviews.

2

Fuzzy inference-based pattern recognition method for sports consumption

Sports marketing utilizes the general principles, methods, and processes of marketing, and is a type of marketing method under the enterprise marketing strategy. In this paper, sports marketing is defined as the marketing activities of an enterprise’s products, services or corporate brand through sports events or sports activities, and the strong appeal and influence of sports events to draw consumers’ attention to the enterprise or its products and services, so as to achieve the purpose of improving the reputation of the enterprise, enhancing the value of the brand, and promoting the consumption and increasing the revenue. The implementation of sports marketing strategies cannot be separated from the identification of customer consumption patterns and behavioral analysis, so this study analyzes sports consumption patterns and behavior based on fuzzy reasoning.

2.1

Fuzzy calculations based on web comments

2.1.1

Fuzzy computation of comment vocabulary

Consumer Internet comments are expressed in the form of natural language, and in the traditional sentiment analysis research on natural language semantic computation, the semantics of words are expressed by the method of exact assignment in classical mathematics, but whether it is the connotation of semantics or its extension, the exact assignment is not able to fully express the meaning of words. The creation of Zadeh’s fuzzy mathematical research led to the idea of assigning intervals or sets to words. Therefore, by combining the fuzzy properties of words, the semantic scribing of natural language is no longer limited to “yes and no”, but is transformed into fuzzy sets by fuzzy affiliation functions.

Naresha et al. gave a method for calculating the fuzzy sentiment intensity of words in conjunction with the study of the semantics of online comment vocabulary [25], based on the degree of praise or blame and the measure of high or low sentiment of the words, which is divided into 9 levels on the fuzzy set, where positive and negative are 4 levels respectively, i.e., S (Small), M (Medium), L (Large), and VL (Vertical), and the measure of non-polarity is Z. Each level corresponds to a fuzzy affiliation function as –VL, –L, –M, –S, Z, +S, +M, +L, +VL, respectively, and the basic fuzzy set is established in the range of [–4,4]. For the selection of the fuzzy affiliation function, based on the rules for the establishment of the affiliation function, we select a symmetric convex fuzzy set, and define a Gaussian-type fuzzy affiliation function on the domain of the argument for representation: (1) $F_{\underset{˜}{w}} (x) = g a u s s m f_{\underset{˜}{w}} (x, σ_{\underset{˜}{w}}, c_{\underset{˜}{w}}) = \exp [\frac{- {(x - c_{\underset{˜}{w}})}^{2}}{2 σ_{\underset{˜}{w}}^{2}}]$

Among them: w ∈ {–VL,–L,–M,–S,Z,+S,+M,+L,+VL}

w represents the nine sentiment levels of the vocabulary, corresponding to the $\underset{˜}{w}$ sentiment levels, $c_{\underset{˜}{w}}$ is the expectation of the affiliation function, and $σ_{\underset{˜}{w}}$ is the standard deviation of the Gaussian affiliation function. The setting of $σ_{\underset{˜}{w}}$ is based on the fact that the greater the polarity of the vocabulary, the higher $σ_{\underset{˜}{w}}$ is, and in order to make the intersection of neighboring affiliation functions take a moderate amount, $σ_{\underset{˜}{w}} = 0.4$ is chosen:

For negative, x ∈ [–4,0], c_–VL = –4, C_–L = –3, c_–M = –2, c_–S = –1;

For positive, x ∈ [0,4], c_+S = 1, c_+M = 2, c_+L = 3, c_+VL = 4.

For words without polarity, a_–A = 0. The larger the absolute value of c_x, the more positive or negative the word is, the more polarized it is, and the higher the emotional intensity of the word.

2.1.2

Semantic fuzzy offset calculation for comment vocabulary

In natural language, words can express certain meanings, but people often modify the original terms through other words to express their own views in the process of communication. The modified semantics are somewhat vague compared to the original, this is due to the flexibility of language. Words that can modify the original term are called linguistic operators, and there are two common types of linguistic operators: tone operators and negation operators. Tone operators are mostly composed of adverbs of degree, which, by modifying the original term, make its single meaning ambiguous and deviate from the original interpretation. Negation operators are mostly composed of negation words, which are able to express the meaning of the negation of the original term [26].

1)

Tone operators

For example, some words are modified by other words such as “very”, “very”, “comparatively”, etc. These modifiers can adjust the semantics of the original word, making it semantically strengthened or weakened. Therefore, tone operators are divided into two categories above the degree of modification of words, which are the centralization operator and the weakening operator. For certain words that can improve the semantic tendency, it is called the centralization operator, and for words that can deviate from the semantic tendency, it is called the weak operator. Algorithmically, V^(λ) is regarded as a tone operator whose modification for the target word A is: $\begin{matrix} V^{(λ)} : F (X) \to F (X) \\ A (x) \to V^{(λ)} A (x) \end{matrix}$ . When λ > 1, V^(λ) is called a centralization operator: when λ < 1, V^(λ) is called a weakening operator.

After screening and organizing the degree words into five categories, namely: trace (VL), low amount (L), medium amount (M), high amount (H), and very high amount (VH), which take the value at [1,5].

For the difference between centralization and weakening of degree words, the extreme, high and medium degree words are regarded as centralization operators, and the low and trace degree words are regarded as weakening operators. Take A-word fuzzification calculation as an example: (2) $F_{\underset{˜}{w}} (A) = g a u s s m f_{\underset{˜}{w}} (x, σ_{\underset{˜}{w}}, c_{\underset{˜}{w}}) = \exp [\frac{- {(x - c_{\underset{˜}{w}})}^{2}}{2 σ_{\underset{˜}{w}}^{2}}]$

After the modification of the word by the tone operator V^(λ), the affiliation function of the word is: (3) $F_{\underset{˜}{w}} (V A) = {[f (x, σ_{\underset{˜}{w}}, c_{\underset{˜}{w}} \pm λ')]}^{λ} = \exp [\frac{- {[x - (c_{\underset{˜}{w}} \pm λ')]}^{2}}{2 σ_{\underset{˜}{w}}^{2}} \cdot λ]$ where $σ_{\underset{˜}{w}} = 0.4$ , w ∈ {–VL,–L,–M,–S,Z,+S,+M,+L,+VL}, and $c_{\underset{˜}{w}}$ are the expectations of the affiliation function corresponding to each sentiment level, and λ and λ′ are the variables of the tone operator, and all are positive real numbers. When w ∈ {+S,+M,+L,+VL}, the function value is shifted by –λ′ units; when w ∈ {–VL,–L,–M,–S}, the function value is shifted by +λ′ units.

2)

Negation operator

Negation can dramatically change the polarity of a word, causing a change in its sentiment. It is not reasonable to take the direct inverse of the semantic change of negative qualifiers and treat words with negative prefixes as MP: (4) $M P (w o r d) = [M P (w o r d) - \frac{M P (w o r d)}{2}]$

Therefore, when blurring words modified by a negative word, a lexical take the opposite and then take half of the word is used to take the example of a A word modified by a negative word N: (5) $F_{\underset{˜}{w}} (N A) = f (x, σ_{\underset{˜}{w}}, - c_{\underset{˜}{w}} / 2) = \exp [\frac{- {(x + c_{\underset{˜}{w}} / 2)}^{2}}{2 σ_{\underset{˜}{w}}^{2}}]$

In practice, however, the more complex change in the emotion of the words is to modify the main word with both the negative word and the adverb of degree, and the expression is mainly categorized into two modes: “negative word-adverb of degree” and “adverb of degree-negative word”. For example, if “pretty” is chosen as the subject word to be modified, “too” is used as the adverb of degree, and “no” is used as the negative word, then the meanings of “not too pretty” and “too unpretty” are quite different. Therefore, when the order of the negation operator and the tone operator is not consistent, the sentiment shift polarity and the degree of shift of the words are not consistent. Taking A word fuzzification calculation as an example, the negation operator of the modifying word is N and the tone operator is V, then the modifications constituting the word are both NVA and VNA: (6) $\begin{matrix} F_{\underset{˜}{w}} (V N A) = {[f (x, σ_{\underset{˜}{w}}, (- c_{\underset{˜}{w}} / 2) \pm λ')]}^{λ} \\ = \exp [\frac{- {[x - (- c_{\underset{˜}{w}} / 2 \pm λ')]}^{2}}{2 σ_{\underset{˜}{w}}^{2}} \cdot λ] \end{matrix}$ (7) $\begin{matrix} {F_{\underset{˜}{w}} (N V A) = [f (x, σ_{\underset{˜}{w}}, - (c_{\underset{˜}{w}} \pm λ') / 4))]}^{λ} \\ = \exp [\frac{- {(x + (c_{\underset{˜}{w}} \pm λ') / 4)}^{2}}{2 σ_{\underset{˜}{w}}^{2}} \cdot λ] \end{matrix}$

By determining the category of modifiers of the target words and calculating based on different modifier combinations, the emotional polarity of the target words and the shift in the degree of emotion can be measured, and ultimately the semantic emotion of the relevant utterance can be calculated.

2.2

Fuzzy reasoning

In order to obtain a logical conclusion, linguistic variables can be used with as input variables to reason according to fuzzy inference rules. The most common form of expression for fuzzy reasoning is IF premise (antecedent), THEN conclusion (consequent). Fuzzy IF-THEN statements are often used to express imprecise reasoning, so this approach plays a more important role for decision-making behavior under uncertainty [27].

Several reasoning algorithms are commonly used in practice, such as Zadeh’s CRI algorithm, Lukasiewicz’s CRI algorithm, Mamdani’s CRI algorithm, etc. Among them, Mamdani’s reasoning is also known as “max-min” reasoning. Among them, Mamdani’s reasoning, also known as “maximum-minimum (max-min)” reasoning, mainly uses F implication relation RM = (A → B) = (A×B), antecedent A*→ consequent B*: (8) $B * (v) = A * (u) \cdot R_{M} (u, v) = V (A * (u) \land A (u) \land B (v))$ (9) $A * (u) = R_{M} (u, v) \cdot B * (v) = V (A (u) \land B * (v) \land B (v))$

2.3

Behavioral Modeling

2.3.1

Constructing fuzzy semantic models

The values taken by linguistic variables are not exact numerical values, but fuzzy sets expressed in fuzzy mathematical language.

Suppose a linguistic variable can be represented by a five-dimensional vector (x,T(x),U,G,M), where x represents the name of the linguistic variable, T(x) represents the linguistic set of the linguistic variable, i.e., the set of names of the linguistic x values, and each value of the linguistic variable corresponds to a fuzzy set over the domain U, G is the syntactic rule for the values of the linguistic variable, and M is the semantic rule for generating the fuzzy set of the linguistic variable’s affiliation function.

Consumers rate the price of an item as low, moderate, or expensive in online reviews. Define “price of the item” as a linguistic variable, or “liking” as a linguistic variable, and so on.

As an example, a consumer can comment on the price of an item in an online review as low, moderate, or expensive. If “item price” is defined as a linguistic variable, then T (item price) may be T(Prices of items) = {Cheap, Moderate, Expensive⋯}.

Define a fuzzy set of the above fuzzy languages such as “inexpensive, moderate, expensive” on Domain U. Setting up domain U = [20,30], it can be roughly considered as “cheap” if it is lower than 20, “expensive” if it is higher than 30, and “moderate” if it is in between. A language value of “inexpensive, moderate, expensive” is assigned to the affiliation function h₁, h₂, h₃, respectively.

Consumer irrational purchasing behavior is divided into 9 levels, each level is expressed by Gaussian affiliation function. The Gaussian affiliation function belongs to the fuzzy control algorithm and is determined by two parameters σ and c: (10) $f (x, σ, c) = c^{\frac{{(x - c)}^{2}}{2 σ^{2}}}$ where parameter b is usually positive and parameter c is used to determine the center of the curve.

2.3.2

Modeling Buying Behavior

Consumers’ irrational purchasing behavior is mainly influenced by three aspects: external inducing factors, internal inducing factors, and the degree of consumer irrationality. Aiming at the online comments of online consumers, the emotional semantic resources in the online comments are mined to study the various factors affecting the irrational behavior of online consumers, so as to analyze and derive the internal and external inducing factors, which prompt consumers to produce irrational purchasing behaviors, and then to judge which one of the nine irrational purchasing degrees of the irrational purchasing impulse behaviors of consumers belongs to. The specific behavioral model is shown in Figure 1.

3

Pattern recognition and behavioral analysis

3.1

Consumption pattern recognition

3.1.1

Individual consumption pattern recognition

In order to verify the feasibility of the method, we take an online review of a user of a sports shopping website as an example for computational analysis. A total of 84 comments on various products and services were obtained by querying the evaluation records of this consumer. These reviews were syntactically analyzed to select opinionated sentences with at least one sentiment word in each sentence. After screening, a total of 115 different emotion words and evaluation words were extracted, which are the key words used to conduct the analysis of consumer irrationality level.

The following is an example of one of the comment statements, “Bought on sale, the price was too attractive, impulsive, feel more satisfied”, to illustrate the process of recognizing consumer consumption patterns from online comments. First, the emotion words in the comment statement are labeled and the parameters of the emotion words are determined. There are three emotional words in this sentence, namely “seductive”, “impulsive”, and “satisfied”, and the corresponding modifiers of these three emotional words are “too”, “NULL”, and “compare” (the word “impulsive” is not modified by an adverb of degree and is marked as “NULL”). Based on the frequency of occurrence of these emotion words in all utterances, parameter h_M(i) (parameter h_M = 1 for the most frequent word among all emotion words) is calculated. The basic sentiment fuzzy sets and modifier operator values for these three sentiment words can be obtained by querying the sentiment vocabulary corpus. The values of each parameter are shown in Table 1.

Table 1.

The parameter of sentiment words calculation

	Affective vocabulary	Modifier Q_M(i)	Basic emotional fuzzy set	h_M(i)	r_M(i)	λ	λ'
M (1)	Tempting	Too	-L	5.33	5.5	6	5.5
M (2)	Impulsive	NULL	-VL	5.44	6.5	6.5	5
M (3)	Satisfactory	Comparatively	-M	5.44	7.5	5.5	6

Secondly, the comprehensive sentiment fuzzy set $\underset{\sim}{M} (i)$ expressed by the consumer through sentiment vocabulary M(i) is calculated, and then the same method is used to obtain the sentiment fuzzy set expressed by the consumer’s sentiment vocabulary, and the comprehensive sentiment fuzzy set of the consumer is calculated $P_{\sim S} (M)$ . Similarly, the evaluation fuzzy set expressed by the consumer’s evaluation vocabulary is obtained, and the comprehensive evaluation fuzzy set of the consumer is calculated $P_{\sim C} (M)$ . Intersection operations are performed on $P_{\sim S} (M) P_{\sim C} (M)$ , and the consumer’s comprehensive sentiment and evaluation fuzzy set $\underset{\sim}{P} (M)$ , and the consumer’s comprehensive sentiment and evaluation fuzzy set is shown in Fig. 2.

Finally, the proximity is calculated in the interval of [–a,a], and the result is: $(\underset{\sim}{P} (M), {\underset{\sim}{A}}_{J}) = \max (\underset{\sim}{P} (M), {\underset{\sim}{A}}_{j}) = 5.9193$ .

Finally, the rationality of the consumer is close to the fuzzy set ${\underset{\sim}{A}}_{9}$ corresponding to the degree of rationality of the category of S₅, the consumer’s degree of rationality mode is “less rational”, the degree of proximity of 5.9193, indicating that the consumer is affected by irrational emotions, belonging to the impulsive consumers.

3.1.2

Group Consumption Pattern Recognition

In turn, 500 more representative online sports consumers are selected. Their online reviews of spoken language characteristics are more obvious, the shopping time and scope are more extensive, and they represent a more representative group of online sports consumers.Calculate the comprehensive emotional value of each of these online consumers and carry out statistics. The distribution of the emotional degree of the consumer group is shown in Figure 3.

The histogram of the distribution of group composite emotions is shown in Figure 4. Use 1-8 to represent these eight intervals respectively. Theoretically, the group of online sports consumers is so large that it is difficult for a single part of individuals to represent the whole group, but it is naturally difficult to select all consumers, so we focus on some more representative individuals.

Using the one-sample normal distribution test method in Matlab, through the fitting test, it is concluded that the distribution of the emotional degree of the sports consumer group approximately obeys the normal distribution under the 95% confidence level, and the probability density curve of the distribution of the emotional degree is shown in Figure 5.

Two similar products from two different merchants on this sports sales platform (soccer star A and star B jerseys, hereafter denoted by C1 and C2, respectively) were further selected, online reviews were extracted, and using the same methodology, it was verified that the degree of emotion generated by the product was also approximated to be normally distributed, and the probability distribution curves of the sums were obtained by fitting as shown in Figure 6.

From the resultant graph, it can be seen that the degree of emotion of the selected group of consumers tends to be irrational. The emergence of more irrational sports consumers in online consumption is consistent with the fact that online consumption has more irrational emotions. Through the comparison of Figure 6, it is found that product C1 has a similar distribution of emotions with the online consumer group, and this type of product causes consumers to purchase relatively rationally, and to some extent lacks the factors that stimulate consumption.

As a similar product, product C2 caused by the average value of the emotional degree curve is higher than the network group, at a higher level, indicating that the product produces more purchasing emotions, has a stronger purchasing attraction.C2 product produces irrational emotions are much higher than C1, through the comparison of the actual purchase volume, the sales volume of C2 is also higher than that of C1, which further verifies the conclusions of this paper. As similar products, the quality of the two products does not differ significantly, while the product’s store carries out a series of promotional offers of marketing strategies, while accumulating a high level of popularity becomes the main reason for the difference between the two.

3.2

Analysis of Sports Consumer Behavior

3.2.1

Analysis of online sports consumption types

The review users of this sports sales platform were selected to analyze the consumption behavior, and the questionnaire survey was taken to analyze the sports consumption level, type and subjective motivation of sports consumers, so as to provide a basis for further proposing and implementing the corresponding sports marketing strategies. A total of 893 questionnaires were distributed, and the valid questionnaires were 842.

Table 2 shows the analysis of the types of online sports consumption of the respondents. Regarding the types of network sports consumption, they are mainly categorized into network sports physical-type consumption and network sports service-type consumption according to the study. According to the statistical analysis, in terms of the type of online sports physical-type consumption, the top three are mainly based on sports shoes, sportswear and sports equipment, accounting for 36%, 30% and 19% respectively, followed by sports protective equipment, accounting for 12%. Sports publishing publications are the least, accounting for 3%. It can be seen that the general consumers in the network sports physical type of consumption, mainly sports equipment, sportswear, sports shoes, mainly physical type of consumption, and mainly for sports necessities consumption.

Table 2.

Analysis of types of online sports consumption of subjects

Variable		Count	Percentage(%)
Online sports physical consumption	Sneaker	303	36%
	Sport clothes	252	30%
	Sport equipment	160	19%
	Sports protection product	101	12%
	Sports publications	26	3%
Online sports service consumption	Gym card	219	26%
	Exercise prescription	152	18%
	Watching sports	143	17%
	Online sports training	126	15%
	Sports tickets	118	14%
	Playground reservation	84	10%

In terms of online sports service consumption types, the top three are mainly sports fitness cards, sports prescriptions, and paying to watch sports events, accounting for 26%, 18%, and 17% respectively. This is followed by online sports training and tickets to sports events, accounting for 15% and 14% respectively. The last is sports venue reservation, accounting for 10%. It can be seen that the current method of choosing sports is more reasonable and scientific compared to the traditional method of choosing sports.

3.2.2

Analysis of online sports consumption levels per quarter

Table 3 shows the analysis of the respondents’ quarterly online sports consumption level. As shown in Table 3, the respondents’ quarterly network sports consumption level is divided into five levels, which are below 200 yuan, 201-300 yuan, 301-400 yuan, 401-500 yuan, and above 501 yuan. The analysis results show that there are 623 people spending less than 200 yuan on online sports per quarter, accounting for 74% of the overall. There are 109 people spending 201-300 yuan on online sports per quarter, accounting for 13% of the overall. There are 42 people spending 301-400 yuan on online sports per quarter, accounting for 5% of the total. 17 people spend 401-500 yuan per quarter on online sports, or 2% of the total. There are 51 people who spend more than 501 yuan on online sports per quarter, accounting for 6% of the total. It can be seen that about 75% of the respondents’ quarterly online sports spending is at a lower level, and there is still a lot of room for improvement.

Table 3.

Quarterly analysis of online sports consumption level

Variable		Count	Percentage (%)
The level of online sports consumption per season	<200 yuan	623	74
	201-300 yuan	109	13
	301-400 yuan	42	5
	401-500 yuan	17	2
	>501 yuan	51	6

3.2.3

Subjective Influences on Online Sports Consumption

The subjective influencing factors affecting sports consumption behavior, i.e. the intrinsic conditions affecting sports consumers’ online purchasing behavior, mainly include sports consumption psychology and sports consumption motivation. Through the analysis of the subjective influencing factors of sports consumption, due to the dependent variable sports consumers network sports consumption level for five classifications, and there is a sequential relationship between the classifications, so for the dependent variable for the case of sub-types of data to choose Logistic regression, so the use of multivariate ordered Logistic regression analysis model for analysis.

With truth-seeking mentality (YES=1, NO=0), herd mentality (YES=1, NO=0), difference-seeking mentality (YES=1, NO=0), comparison mentality (YES=1, NO=0), affordable price (YES=1, NO=0), product diversity (YES=1, NO=0), ease of purchasing (YES=1, NO=0), saving time (YES=1, NO=0), and diversity of payment methods (YES=1, NO=0) as independent variables, and quarterly online sports consumption (below 200 yuan = 1, 201-300 yuan = 2, 301-400 yuan = 3, 401-500 yuan = 4, 501 yuan and above = 5) as dependent variables, to establish a multivariate ordered logistic regression analysis model, and the results are shown in Table 4.

Table 4.

Logistic regression of subjective factors affecting sports consumption

Variable	Estimate	SE	Wald	Df	Sig.	95% confidence interval
Variable	Estimate	SE	Wald	Df	Sig.	Upper limit	Lower limit
Sports consumption=1	5.825	0.640	121.677	1	0	4.869	6.888
Sports consumption=2	6.846	0.620	156.948	1	0	5.813	8.227
Sports consumption=3	7.817	0.582	180.851	1	0	6.765	8.756
Sports consumption=4	7.863	0.583	190.324	1	0	6.869	9.194
Affordability	1.683	0.331	21.775	1	0.04	1.013	2.306
Product diversity	0.893	0.238	12.432	1	0	0.242	1.403
Convenience of purchase	0.658	0.315	5.870	1	0.013	0.030	1.166
Save time	0.355	0.398	5.187	1	0.021	0.187	0.980
Payoff diversity	-0.357	0.223	1.638	1	0.232	-0.797	0.240
Practical psychology	1.326	0.198	14.832	1	0	0.434	2.095
Conformity mentality	1.090	0.263	33.174	1	0	0.749	1.587
Mindset of chasing difference	1.348	0.282	18.325	1	0	0.724	1.584
Mindset of comparison	1.521	0.443	23.630	1	0	0.927	2.213

According to Table 4, it can be seen that the consumer psychology (truth-seeking psychology, herd mentality, differentiation mentality, comparison mentality) all have a significant effect on influencing sports consumers’ online sports consumption (P<0.05). With the rapid development of mobile Internet technology, consumers’ self-control is challenged when consuming sports. On the one hand, it is affected by market factors. The market environment is more complex, and consumers lack experience in online sports consumption and have not formed a stable concept of online sports consumption, which leads to the emergence of herd and comparison consumption psychology. Some consumers are difficult to formulate a reasonable network sports consumption plan according to their own economic conditions and actual needs, and lack of subjectivity in network sports consumption, resulting in blind and irrational herd consumption concepts. On the other hand, influenced by the surrounding environment, consumers like to compare consumption with friends and classmates in their daily behavior to satisfy their own psychological balance and gain their recognition to satisfy their own vanity, without combining their own economic conditions to rationally consume, and imitating high-grade consumption to appear the psychology of comparison consumption.

4

Implementation strategies for sports marketing based on the results of the analysis

From the results of pattern recognition and behavioral analysis, it can be seen that irrational consumption occupies a considerable portion of the surveyed sports consumers, and factors such as the herd mentality, the psychology of seeking differences, and the psychology of comparison are easy to guide sports consumers to impulsive consumption. Aiming at this characteristic, this paper proposes a variety of sports marketing implementation strategies to stimulate the level of sports consumption.

4.1

Focus on product content

No matter what kind of marketing strategy is used, is to improve product sales, so that the majority of consumers to consume, but the product must withstand product viewing considerations, the first point is the product quality, followed by product connotation, especially product connotation, only the enterprise to the consumer’s values, social outlook, consumerism and product connotation, so that its brand image has the spirit of humanism, so as to expand the marketing Channels, such as Li Ning clothing enterprises, it is transmitted by the human spirit will have the breath of the times, under the development of sports marketing strategy to impress consumers and enrich the content of enterprise products. For sports marketing, the product needs to highlight the emphasis on sportsmanship, so that when compared with other items of the same type, it will be more competitive in the market, and the product connotation also shows the side of the enterprise’s image, thus promoting the maximization of corporate profits.

4.2

Establish a correct marketing concept

In the development process of sports marketing, we must first start from the consumer side, correctly deal with the relationship between consumers and sellers, and consider the future development of the enterprise, that is to say, the enterprise in the operation of sports marketing at the beginning of the product and the consumer’s feelings, social interests and other aspects of the organic unity of the product and the implementation of the implementation of the implementation of the present day’s enterprise sports marketing, there are many enterprises are only concerned about their products and In today’s corporate sports marketing, there are many companies are only concerned about how to make their products and brands bigger and stronger, using various means to obtain the growth of sales figures, and for the community, what kind of impact on the audience is generally difficult to take into account, only from their own interests.

4.3

Closely linking sports culture with corporate brand image

For consumers, things with cultural feelings are more attractive, more flavor, so in the sports marketing process, the need for sports culture and corporate brand image closely linked to the seller and buyer based on sports activities and produce a common focus, the formation of psychological resonance, such as to find an effective connection between sports celebrities and corporate brands, because not any sports stars are able to Competent in every product promotion, and celebrities said to the public personality, words and deeds, image also need to match with the image of the corporate brand, thus increasing the advertising effect and achieve win-win situation.

4.4

Exploiting extreme tendencies in collective decision-making

When consumers make decisions, those who were inclined to take risks when making decisions individually will be more adventurous after a group discussion, while those who were inclined to be conservative when making decisions individually will be more conservative after a group discussion. Therefore, when selling to groups of consumers, such as sports goods, it is more effective to bring people who had the intention to buy together again to sell. The people who have no intention to buy should be sold individually to avoid them becoming more conservative after the collective discussion.

5

Conclusion

In this paper, a fuzzy calculation was carried out based on the network comments to identify the pattern of sports consumption, and the behavior of sports consumption was identified. The proximity degree of the selected individual consumer is 5.9193, which recognizes that the consumer’s sports consumption pattern is close to “impulsive consumption”. The consumption pattern of 500 online sports consumers was identified through their comments. The emotional level of this consumer group tends to be irrational, and the mean value of the emotional level curve of C2 products is higher due to the marketing strategy of promotion optimization adopted by the merchants.

After verifying the feasibility of recognizing consumption patterns in this paper, consumption behavior analysis is carried out.The top three sports consumed types are sports shoes (36%), sports clothing (30%), and sports equipment (19%). Sports consumption is mostly centered around sports necessities.There is still room for improvement in the sports consumption level of most consumers. In addition, the psychology of following the crowd, seeking differences and comparing significantly affects sports consumption (p<0.05).

On this basis this paper proposes that marketing strategies should be implemented in terms of product connotation, marketing concept, sports culture and collective decision-making.

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

Pattern Identification and Behavioral Analysis of Sports Marketing Strategy Implementation in Big Data Environment

Li Zhang

Ke Sun

Jinjin Ren

Publicado en línea: 21 mar 2025

Recibido: 29 oct 2024

Aceptado: 01 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0646

Palabras claveFuzzy computing, Sports marketing strategy, Pattern recognition, Behavioral analysis, Big data

© 2025 Li Zhang et al., published by Sciendo

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

Palabras clave
Fuzzy computing, Sports marketing strategy, Pattern recognition, Behavioral analysis, Big data