A SYSTEM AND METHOD FOR ANALYZE BUYING BEHAVIOR OF CUSTOMER

Abstract

ABSTRACT A SYSTEM AND METHOD FOR ANALYZE BUYING BEHAVIOR OF CUSTOMER The present disclosure relates to the field of marketing analytics and consumer behavior analysis. More specifically, the present disclosure relates to the system and method for determining a product preference score for a consumer using an artificial neural network method. Also, the present disclosure provides more effective, empathetic, and consumer-centric approaches that enhance both customer satisfaction and business outcomes. Further, the method includes receiving data associated with a set of consumer behavior. Furthermore, the method includes determining a product preference score for the consumer using a artificial neural network model based on the quantified value of each generic attribute of the set of generic attributes of the product, the product category-specific attributes of the product, and the received data associated with the set of consumer behaviour.

Specification

Description:TECHNICAL FIELD:
The present disclosure relates to the field of marketing analytics and consumer behavior analysis. More specifically, the present disclosure relates to the system and method for determining a product preference score for a consumer using an artificial neural networks method. Further the system and method which is integrating psychological insights into marketing strategies to better understand and meet consumer needs. Marketers can create more effective, empathetic, and consumer-centric approaches that enhance both customer satisfaction and business outcomes when determining the product preference for a consumer.

BACKGROUND:
The background description includes information that may help understand the current invention. It is not an admission that any of the data presented here is previous knowledge or pertinent to the invention currently being claimed, nor is it an admission that any publication directly or implicitly referred to herein is prior knowledge.

Eben Harrell (2019), senior editor at Harvard Business Review, the field of neuromarketing-sometimes known as consumer neuroscience-studies the brain to predict and potentially influence consumer behaviour and decision-making. Over the past five years, several groundbreaking studies have demonstrated its potential. Nobel Laureate Francis Crick referred to this concept as the "astonishing hypothesis," the idea that all human feelings, thoughts, actions, and even consciousness itself are merely the products of neural activity in the brain. For marketers, the promise of this idea lies in the potential of neurobiology to reduce the uncertainty and guesswork that traditionally hinder efforts to understand consumer behaviour. Previously considered an extravagant "frontier science," neuromarketing has gained credibility in recent years, bolstered by several groundbreaking studies that demonstrate its value for marketers.

Cherubino et al. (2019) described how the possibility of getting inside the minds of customers has given companies the opportunity to better understand the consumer behaviour that underlies the decision-making process. Neuromarketing techniques have enabled businesses to discover and target the real needs, desires, and wants of individuals. The authors discuss key elements such as heart rate, galvanic skin response, eye tracking, reaction time tests, and facial expressions, which are proving essential in understanding how products entice consumers to make purchasing decisions.

Etzold et al. (2019) explained that consumer behaviour involves examining individuals, groups, or organizations, along with all the behaviours related to the acquisition, use, and disposal of products and services. It is the study of how a consumer's emotions, attitudes, and preferences influence their purchasing decisions. Market research, often based on customers' subconscious emotions, is conducted to improve the efficiency and effectiveness of marketing campaigns. For example, eye-tracking research on consumer cognition helps marketers better understand human behavior, making eye tracking a rapidly growing multidisciplinary field that integrates electronics, neurology, and cognitive science to investigate how people solve problems and make choices.

Wu et al. (2021) focused their study on consumer cognition, particularly reflecting on how consumer behaviour is influenced by visual sensors. They employed the Tobii eye tracker to examine human eye movements, exploring the eye-mind connection and how eye motion data can express people's activities. Their findings established a significant link between human cognition and eye movements toward products.

Pantaewan et al. (2012), Artificial Neural Networks (ANNs) are widely used computational methods that help solve complex problems by simulating animal brain processes in a simplified manner.

Svajone Bekesiene et al. (2021) described ANN approaches, such as Perceptron-type neural networks (PTNNs), which consist of artificial neurons (nodes) that serve as information processing units. These artificial neurons are organized into layers and interconnected by synaptic weights (connections). Through this information processing method, neurons can screen and transmit data in a structured, supervised manner to construct an analytical model capable of classifying the stored data. Typically, ANNs are designed as three-layer network models of interconnected artificial neurons, comprising an input layer, a hidden layer, and an output layer. It is also possible for researchers to include one or more hidden layers between the input and output layers. Furthermore, while neurons within the same layer do not have interconnections, each neuron can still be linked to neurons in the subsequent layer.

The system and method for calculating a consumer's buying behaviour preference score. Understanding consumer trends and behaviour is important to today's users and sellers in order to better marketing tactics, streamline inventory control, and boost overall company success. Users and sellers can customize their products, modify their marketing tactics, optimize their prices, and enhance the entire consumer experience by using the score that the product preference determination algorithm has generated. The consumer buying behaviour preference determination system gives users and sellers insights that help them remain flexible in the face of shifting market conditions. Their ability to adapt enables them to react proactively to changes in the market, competition, and consumer demand.

Thus, there is a need for a system and method which is integrating psychological insights into marketing strategies to better understand and meet consumer needs. Marketers can create more effective, empathetic, and consumer-centric approaches that enhance both customer satisfaction and business outcomes when determining the product preference for a consumer.

OBJECTIVES:
An objects of the present disclosure relates to a novel system for analysing consumer behaviour.

Another object of the present disclosure relates to a method for marketing analytics and consumer behaviour analysis.

Yet another object of the present disclosure relates to a system and method for determining a product preference score for a consumer using a aartificial nneural nnetworks method.

Further object of the present disclosure relates to the system and method which is integrating psychological insights into marketing strategies to better understand and meet consumer needs.
Furthermore, object of the present disclosure relates to the system to provide marketers can create more effective, empathetic, and consumer-centric approaches that enhance both customer satisfaction and business outcomes when determining the product preference for a consumer.

SUMMARY:
This summary is provided to introduce a selection of concepts, in a simplified format, that are further described in the detailed description of the invention. This summary is neither intended to identify key or essential inventive concepts of the invention and nor is it intended for determining the scope of the invention.

An aspect of present disclosure relates to a system (100) for determining the buying behavior of products, the system (100) comprising:
i. a neuroticism trait (101);
ii. a neuroticism trait variable (102);
iii. an artificial neural network (103); and
iv. a buying behavior module (104);
wherein the buying behavior determination module (104) determines the buying behaviors of the consumer by applying an artificial neural network (ANN) (103) based on the value of each variable of the neuroticism traits (101) and the data associated with the consumer behaviors.

Another aspect of present disclosure relates a method for determining the buying behavior of product, the system (100) comprising:
i. collecting and analyzing data associated with a set of consumer behaviors, wherein the consumer attributes neuroticism traits;
ii. determining a product preference score for the consumer using an artificial neural network model (103) based on the value of each neuroticism trait variable (103) of the product; and
iii. the data associated with the set of consumer behaviors.

As illustrated in the accompanying FIGURES, the features and advantages of the subject matter hereof will be made clearer by the extensive a description of a few selected implementations that follows. A person who is reasonably competent in the art will be aware that the disclosed subject matter can be altered in a variety of ways without crossing any boundaries. So, it is fair to say that the photographs and the explanation are illustrative.

BRIEF DESCRIPTION OF DRAWINGS:
The present subject matter will now be described in detail with reference to the drawings, which are provided as illustrative examples of the subject matter to enable those skilled in the art to practice the subject matter. It will be noted that throughout the appended drawings, features are identified by like reference numerals. Notably, the FIGUREs and examples are not meant to limit the scope of the present subject matter to a single embodiment, but other embodiments are possible by way of interchange of some or all of the described or illustrated elements and, further, wherein

1. Figure 1. Relationship between Consumers' neuroticism traits and buying behaviour mediating through artificial neural networks.
2. Figure 2. Conceptual framework of consumers' neuroticism traits on buying behaviour using ANN model.
3. Figure 3 Hidden Layer Activation Function: Hyperbolic Tangent Output Layer Activation Function.
4. Figure 4.ANN model's Predicted Pseudo-Probability presented by box plot diagram for the four Buying Behaviour categories.
5. Figure 5. ROC curve for the ANN model.
6. Figure 6 (a). Model Performance Measurement: Cumulative Gains that are the Illustrations of Accurate Classifications Gained by the ANN model.
7. Figure 6 (b). Model Performance Measurement: Lift Chart Showing Model Performance in a Portion of the Population.
8. Figure 7. Normalized importance by the Artificial Neural Network (ANN) model on Neuroticism Traits variables.


DETAILED DESCRIPTION:
For the purpose of promoting an understanding of the principles of the invention, reference will now be made to the embodiment illustrated in the drawings and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of the invention is thereby intended, such alterations and further modifications in the illustrated system, and such further applications of the principles of the invention as illustrated therein being contemplated as would normally occur to one skilled in the art to which the invention relates. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skilled in the art to which this invention belongs. The system, methods, and examples provided herein are illustrative only and not intended to be limiting

All technical and scientific words used herein, unless otherwise defined, have the same meaning as commonly known by one having ordinary ability in the relevant field. The words "a" and "an" may," "can," "could," or "might" be included or have a characteristic designate one or more (i.e., at least one) of the article's grammatical objects. Unless otherwise stated, weight is used to determine all percentages and ratios. All percentages and ratios are calculated based on the total final composition unless otherwise indicated.

As used herein, whether in a transitional phase or the body of a claim, the terms "comprise(s)" and "comprising" are to be interpreted as having an open-ended meaning. That is, the terms are to be interpreted synonymously with the phrases "having at least" or "including at least". When used in the context of a process, the term "comprising" means that the process includes at least the recited steps but may include additional steps. When used in the context of a composition, the term "comprising" means that the composition includes at least the recited features or components but may also include additional features or components.

The phrases "in an embodiment," "according to one embodiment," and the like generally mean the particular feature, structure, or characteristic following the phrase is included in at least one embodiment of the present disclosure and may be included in more than one embodiment of the present disclosure. Importantly, such phrases do not necessarily refer to the same embodiment.

It will be appreciated by those of ordinary skill in the art that the diagrams, schematics, illustrations, and the like represent conceptual views or processes. illustrating systems and methods embodying this invention. The functions of the various elements shown in the figures may be provided through the use of dedicated hardware as well as hardware capable of executing associated software. Similarly, any switches shown in the figures are conceptual only. Their function may be carried out through the operation of program logic, through dedicated logic, through the interaction of program control and dedicated logic, or even manually, the particular technique being selectable by the entity implementing this invention. Those of ordinary skill in the art further understand that the exemplary hardware, software, processes, methods, and/or operating systems described herein are for illustrative purposes and, thus, are not intended to be limited to any particular name.

All composition described herein can be performed in suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., "such as") provided with respect to certain embodiments herein is intended merely to better illuminate the disclosure and does not pose a limitation on the scope of the invention otherwise claimed. No language in the specification should be construed as indicating any non-claimed element essential to the practice of the disclosure.

The present disclosure overcomes the aforesaid drawbacks of the prior art and other objects, features, and advantages of the present disclosure will now be described in greater detail. Also, the following description includes various specific details and, is to be regarded as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that without departing from the scope and spirit of the present disclosure and its various embodiments there may be any number of changes and modifications described herein.

The numerical values given for various physical parameters, dimensions and quantities are only approximate values and it is envisaged that the values higher than the numerical value assigned to the physical parameters, dimensions and quantities fall within the scope of the disclosure unless there is a statement in the specification to the contrary.

An aspect of the present disclosure provides, a system (100) for determining the buying behavior of products, the system (100) comprising:
v. a neuroticism trait (101);
vi. a neuroticism trait variable (102);
vii. an artificial neural network (103); and
viii. a buying behavior module (104);
wherein the buying behavior determination module determines the buying behaviors of the consumer (104) by applying an artificial neural network (ANN) (103) based on the value of each variable of the neuroticism traits and the data associated with the consumer behaviors.

An embodiment of the present disclosure provides the neuroticism traits (101) comprise anxious, angry, depressed, self-conscious, impulsive, and vulnerable.

In an embodiment of the present disclosure provides the neuroticism trait variables (102) labeled NEU-1 to NEU-9.

Another embodiment of the present disclosure provides the system (100) as claimed in claim 1, wherein the artificial neural network (ANN) (103) is a multilayer perceptron (MLP) neural network, is to identify the neuroticism traits that influence the buying behavior of products by analyzing self-reported data.

Yet another embodiment of the present disclosure provides the multilayer perceptron (MLP) neural network was automatically designed with three nodes in the hidden layer and four nodes in the output layer to represent the dependent variable, buying behavior.

In yet another embodiment of the present disclosure provides the data were collected through a multi-stage random sampling method, followed by a survey to generate data for training and testing the ANN models.

Further embodiment of the present disclosure provides a method for determining the buying behavior of product, the system (100) comprising:
iv. collecting and analyzing data associated with a set of consumer behaviors, wherein the consumer attributes neuroticism traits;
v. determining a product preference score for the consumer using an artificial neural network model based on the value of each neuroticism trait variable of the product; and
vi. the data associated with the set of consumer behaviors.

Figure 1 illustrates an exemplary environment 100 having a system 101 for neuroticism traits (include anxious, angry, depressed, self-conscious, impulsive, and vulnerable) for a consumer, in accordance with an embodiment of the present disclosure. The system 100 may include a neuroticism variable 102 includes NEU1 to NEU 9, an artificial neural network 103, a buying behaviour 104 (may include Personalized Marketing Strategies, Enhanced Consumer Segmentation, Improved Product Development, Effective Communication Channels, Strategic Customer Relationship Management (CRM), Ethical Considerations:), a database, and a product preference system. The system 100 may, without departing from the scope of the disclosure, either be implemented on the consumer or be communicatively coupled to the user device.

Sample and Data Collection
The present disclosure focuses on predicting how consumers' neuroticism traits influence buying behaviour for FMCG products. In this context, the research evaluates the importance and relevance of current conditions using a descriptive research approach. The aim is to describe the current impact of consumers' neuroticism traits on FMCG buying behaviour, making the descriptive research method appropriate for this objective. Consequently, the study selected 600 samples using a multi-stage random sampling method across the Pondicherry territory. The samples, which were normally distributed, were collected from four areas: Pondicherry, Karaikal, Mahe, and Yanam, with 150 samples from each location, totalling 600 samples enrolled in the study.

Interpretation of Research Results
The IBM SPSS Statistics 21v was used to construct the neural network models and to examinate their precision. The investigations according the ANN models' training, construction and testing stages are presented below. The analysis starts with the descriptive statistics of the nine Neuroticism Traits variables, providing the frequency rates such as mean, standard deviation, Variance, skewness and Kurtosis (table-1). The results are shown below

Table -1. Descriptive Statistics for Neuroticism Traits Variables
Variable N Min Max Mean S. D Variance Skewness Kurtosis
Statistic S.E Statistic S.E
NEU-1 600 1.00 5.00 3.026 1.346 1.812 0.030 0.100 -1.199 0199
NEU-2 600 1.00 5.00 3.295 1.280 1.641 -0.211 0.100 -1.143 0.199
NEU-3 600 1.00 5.00 3.380 1.370 1.879 -0.317 0.100 -1.189 0.199
NEU-4 600 1.00 5.00 3.385 1.351 1.826 -0.353 0.100 -1.170 0.199
NEU-5 600 1.00 5.00 2.085 1.050 1.103 0.966 0.100 0.530 0.199
NEU-6 600 1.00 5.00 3.183 1.325 1.756 -0.133 0.100 -1.172 0.199
NEU-7 600 1.00 5.00 2.816 1.359 1.849 0.187 0.100 -1.161 0.199
NEU-8 600 1.00 5.00 2.695 1.315 1.732 0.285 0.100 -1.073 0.199
NEU-9 600 1.00 5.00 2.313 1.121 1.257 0.587 0.100 -0.344 0.199

Table 2. Pearson's Correlation Relationships Between Research Variables
NEU-1 NEU-2 NEU-3 NEU-4 NEU-5 NEU-6 NEU-7 NEU-8 NEU-9
NEU-1 1 0.336** 0.349** 0.290** 0.158** 0.050 0.061 -0.001 0.051
NEU-2 0.336** 1 0.628** 0.386** 0.115** -0.038 0.188** 0.052 -0.011
NEU-3 0.349** 0.628** 1 0.437** 0.056 -0.063 0.166** 0.038 0.108**
NEU-4 0.290** 0.386** 0.437** 1 0.160** -0.008 0.072 0.120** 0.079
NEU-5 0.158** 0.115** 0.056 0.160** 1 0.298** 0.026 0.274** 0.125**
NEU-6 0.050 -0.038 -0.063 -0.008 0.298** 1 0.320** 0.292** 0.199**
NEU-7 0.061 0.188** 0.166** 0.072 0.026 0.320** 1 0.265** 0.146**
NEU-8 -0.001 0.052 0.038 0.120** 0.274** 0.292** 0.265** 1 0.155**
NEU-9 0.051 -0.011 0.108** 0.079 0.125** 0.199** 0.146** .155** 1
**. Correlation is significant at the 0.01 level (2-tailed).

Additionally, Table 2 presents the Pearson's correlation coefficient, which was used to evaluate the relationships between the constructed variables. This analysis identified significant correlations between the neuroticism traits variables at the 0.01 level (two-tailed). The highest significant correlation was found between the variables "I see myself as someone who is not easily stressed" (NEU-2) and "I see myself as someone who does not worry a lot" (NEU-3) (r=0.628, p < 0.001). The next highest significant correlation was between "I see myself as someone who does not worry a lot" (NEU-3) and "I see myself as someone who is emotionally stable and doesn't get upset easily" (NEU-4) (r=0.437, p < 0.001). Following that, a significant correlation was also observed between "I see myself as someone who is not easily stressed" (NEU-2) and "I see myself as someone who is emotionally stable and doesn't get upset easily" (NEU-4) (r=0.386, p < 0.001).

Results of Case Processing
This analysis examines whether the Multi-Layer Perceptron (MLP) neural network can identify the key neuroticism traits that influence the buying behavior of FMCG products by analyzing self-reported data. Table 3 presents the number of neurons in each layer, along with nine independent variables labeled NUE-1 to NUE-9. The MLP neural network was automatically designed with three nodes in the hidden layer and four nodes in the output layer to represent the dependent variable, buying behavior, categorized as 2.00 = Disagree, 3.00 = Neutral, 4.00 = Agree, and 5.00 = Strongly Agree. Different functions were applied to different layers: the hyperbolic tangent function was used as the activation function for the hidden layer, while the SoftMax function was applied to the output layer. The model's validation, specifically when using the SoftMax function, was evaluated using cross-entropy as the error function.

Table 3. Network Information for Case Processing and Summary for Designed Models
Layer Description Variable Description
Layer Number of Units Activation Function NEU-1, NEU-2, NEU-3, NEU-4, NEU-5, NEU-6, NEU-7, NEU-8, NEU-9,
Dependent variable-Buying Behaviour,
2.00 = Disagree, 3.00 = Neutral, 4.00 = Agree, 5= 00 Strongly Agree
Input 9 ANN-Hyperbolic Tangent SoftMax
Hiddena 3
Output 4
a. Excluding the bias unit
Summary for Designed Models
Layer Description ANN
Training Cross Entropy Error 348.570
Percent Incorrect Predictions 38.5%
Stopping Rule Used 1 consecutive step(s) with no decrease in errora
Training Time 0:00:00.30
Testing Cross Entropy Error 141.143
Percent Incorrect Predictions 41.5%
a. Error computations are based on the testing sample, Dependent Variable: Buying Behaviour

Table 4. Survey Sample Classification of the ANN model
Sample Observed Predicted Buying Behaviour
2.00 3.00 4.00 5.00 Percent Correct
Training 2.00 Disagree 0 14 4 0 0.0%
3.00 Neutral 0 157 61 0 72.0%
4.00 Agree 0 77 107 0 58.2%
5.00 Strongly Agree 0 0 9 0 0.0%
Overall Percent 0.0% 57.8% 42.2% 0.0% 61.5%
Testing 2.00 Disagree 0 8 0 0 0.0%
3.00 Neutral 0 52 25 0 67.5%
4.00 Agree 0 36 48 0 57.1%
5.00 Strongly Agree 0 0 2 0 0.0%
Overall Percent 0.0% 56.1% 43.9% 0.0% 58.5%
Dependent variable: Buying Behaviour: 2.00 = Disagree, 3.00 = Neutral, 4.00 = Agree, 5.00=Strongly Agree

The IBM SPSS 21v program was used to predict the buying behavior of FMCG products, categorized as 2.00 = Disagree, 3.00 = Neutral, 4.00 = Agree, and 5.00 = Strongly Agree, using nine variables. The network diagram, shown in Figure 3, illustrates the model with nine input nodes, three hidden nodes, and four output nodes, corresponding to the four categories of buying behavior. A summary of the designed models, presented in Table 3, provides information on the results of the training and testing samples.

Cross-entropy error was used as the error function for both the training and testing samples during the neural network training stage. The ANN model was identified with a cross-entropy error value of 348.57, indicating its capability to predict buying behavior of FMCG products based on neuroticism traits. According to the research results, the percentages of incorrect predictions made by the ANN model were 38.5% for the training sample and 41.5% for the testing sample. The training procedure continued until a consecutive step showed no decrease in the error function.

Table 4 provides a description of the ANN model's case classification for the buying behavior of FMCG as a categorical dependent variable, by partition and in total. The forecast outcome by the ANN model for each case was considered correct if the predicted probability was greater than 0.5. As shown in Table 4, the ANN model correctly classified 61.5% of the training cases and 58.5% of the testing cases overall.
Table-5 Parameter Estimates

Predictor
Predicted
Hidden Layer 1 Output Layer
H(1:1) H(1:2) H(1:3) Buying
Behaviour
= 2.00 Buying
Behaviour
= 3.00 Buying
Behaviour
= 4.00 Buying
Behaviour
=5.00
Input Layer (Bias) -0.980 -0.521 -0.487
NEU-1 0.392 -0.037 -0.113
NEU-2 -0.161 0.103 0.001
NEU-3 -0.153 0.084 -0.131
NEU-4 0.332 0.424 0.043
NEU-5 0.114 0.060 -0.704
NEU-6 0.198 -0.463 -0.770
NEU-7 1.031 -0.059 -0.310
NEU-8 0.123 0.003 -0.149
NEU-9 -0.109 -0.250 -0.286
Hidden Layer 1 (Bias) -0.0781 1.229 1.267 -1.439
H(1:1) -0.326 -0.700 0.151 1.474
H(1:2) -0.196 0.178 -0.501 0.362
H(1:3) 0.671 -0.119 -0.323 -0.575

Additionally, the analysis conducted using IBM SPSS 21v software presented the predicted pseudo-probabilities of the ANN model for the four categories of the Buying Behaviour variable in a box-plot diagram. This specific graph illustrated the predictions for each of the four categories of Buying Behaviour for FMCG products separately. The box plots categorized the predicted pseudo-probabilities based on the entire analyzed dataset. According to the established rule, for each box plot in each category, values above 0.5 indicate correct predictions.

The box plot developed from the ANN model showed the predicted probability for each of the four categories: 2.00 = Disagree, 3.00 = Neutral, 4.00 = Agree, and 5.00 = Strongly Agree. A detailed analysis of the diagram should start from the left side, representing the Disagree category. The first box plot from the left corresponded to the Neutral category, while the second box plot demonstrated the probability for the Neutral category being incorrectly classified in the Strongly Agree category, even though it was actually in the Neutral category. The third box plot showed a zero value for outcomes indicative of the Strongly Agree category, and the fourth box plot similarly showed a zero value for outcomes indicative of the Disagree category.

Furthermore, the probabilities predicted by the ANN model for all four categories revealed that the probability for Agree was close to one, while the Agree probability in both the Disagree and Strongly Agree categories was zero. These findings suggest that the ANN model effectively classified the cases, as shown in Figure 4.

Table 6. Area under the Curve (AUC)
Dependent Variable Categories Area
Buying Behaviour 2.00 Disagree 0.698
3.00 Neutral 0.698
4.00 Agree 0.693
5.00 Strongly Agree 0.954

The area under the curve (AUC), used as a dimensional index, helped summarize the overall accuracy of the ROC curves for the designed ANN model across the four Buying Behaviour categories. This information is crucial as it provides a meaningful interpretation for researchers. The AUC, presented in Table 6, can be understood as the probability that a randomly selected individual will be correctly rated or ranked as having a Strongly Agree buying behaviour based on their neuroticism traits. This interpretation is derived from non-parametric Mann-Whitney U statistics, which are used in computing the AUC. Additionally, the maximum AUC value of 0.954 (as shown in Table 6) indicates that the ANN model was highly effective in predicting Strongly Agree buying behaviour through the measured neuroticism traits.

Figure 5 presents the sensitivity and specificity diagram for Buying Behaviour categories (2.00 = Disagree, 3.00 = Neutral, 4.00 = Agree, 5.00 = Strongly Agree), based on both the training and testing data. The 45-degree line, running from the upper right to the lower left of the chart, represents the scenario of randomly guessing the category. The further the curve deviates from this 45-degree reference line, the more accurate the classification.

The area under the curve (AUC) was measured, and the best result, 0.954, was achieved for the Strongly Agree category in predicting Buying Behavior based on Neuroticism Traits. The AUC values for the other categories were 0.693 for Agree, 0.698 for Neutral, and 0.698 for Disagree. Figure 5 illustrates these results, showing the performance of the ANN model under the ROC curve.

Additionally, Figure 6a shows the cumulative gains chart, illustrating the precise classifications achieved by the ANN model compared to the correct classifications that could be expected by chance (i.e., without using the model). The gain chart demonstrates that for the Strongly Agree category, as indicated by the fourth point on the curve at (10%, 100%), if the network scores and sorts all data by the predicted pseudo-probability of Strongly Agree, the top 10% of cases would include approximately 100% of all Strongly Agree cases. This implies that it is not necessary to select 100% of the scored data to identify all Strongly Agree samples in the dataset. Thus, the gain chart reflects the effectiveness of the ANN model's classification.

The baseline and curve position in the gain chart allow us to assess the model's performance. Figure 6a reveals that the constructed model shows a high overall gain and demonstrates excellent performance for predicting Strongly Agree buying behavior through neuroticism traits. However, it also indicates that the Neutral and Agree categories are less accurately predicted.

Figure 6b presents the lift diagram, which, along with the gain chart, provides graphical support for evaluating the performance of classification models. Unlike the confusion matrix, which assesses models across the entire population, gain and lift diagrams focus on a subset of the population. The lift diagram offers a clear view of the model's advantages. Measures from the gain chart were used to compute the lift aspect (i.e., the benefit). For the Strongly Agree category, the lift at 100% was calculated as 100% / 8% = 12.5.

Table -7. Independent Variable Importance
Neuroticism Traits variables Importance Normalized Importance
NEU-1 The time of Purchase I see myself as someone who is not miserable 0.101 33.6%
NEU-2 The time of Purchase I see myself as someone who can be not stressed 0.039 13.0%
NEU-3 The time of Purchase I see myself as someone who not worries a lot 0.037 12.3%
NEU-4 The time of Purchase I see myself as someone who is emotionally stable and doesn't get upset easily 0.116 38.5%
NEU-5 The time of Purchase I see myself as someone who can be secure 0.133 44.3%
NEU-6 The time of Purchase I see myself as someone who is sometimes not shy and inhibited 0.157 52.2%
NEU-7 The time of Purchase I see myself as someone who gets not nervous easily 0.301 100.0%
NEU-8 The time of Purchase I see myself as someone who does remain calm in high-pressure situations 0.046 15.2%
NEU-9 The time of Purchase I see myself as someone who does remain calm in tense situations 0.070 23.4%

The impact of each independent variable, in terms of relative and normalized importance, as identified in the designed ANN models, is displayed in Table 7. Figure 7 presents charts that illustrate the importance of consumers' neuroticism traits for predicting FMCG buying behaviour.

When analyzing the results, it is noted that NEU-7 ("The time of purchase, I see myself as someone who does not get nervous easily") had the highest importance among all predictors, with a normalized importance of 100% in the neural network models. Other predictors with significant importance included NEU-6 ("The time of purchase, I see myself as someone who is sometimes not shy and inhibited") with a normalized importance of 52.2%, and NEU-5 ("The time of purchase, I see myself as someone who can be secure") with a normalized importance of 44.3%. Conversely, NEU-2 ("The time of purchase, I see myself as someone who is not easily stressed") and NEU-3 ("The time of purchase, I see myself as someone who does not worry a lot") had lower importance, with normalized values of 13% and 12.3%, respectively.
Using non-linearity in the study of consumer buying behavior helps address the complex question of how neuroticism traits influence consumers. This approach is particularly useful when examining diverse social and physical behaviors.

The study results, as shown in Table 4, indicate that the most effective model was the Multi-Layer Perceptron Neural Network (MLPNN) with 9 input neurons, 3 hidden neurons, and 4 output factors. The model, trained with a hyperbolic tangent activation function and a SoftMax function in the output layer, achieved the lowest cross-entropy error values (348.57 for training and 141.14 for testing). This suggests that the ANN model with these functions provided a high validation result and effectively predicted consumer buying behavior based on neuroticism traits.

The ANN model, trained using the back-propagation algorithm, identified the predictor "time of purchase, I see myself as someone who does not get nervous easily" as having the highest importance, with a normalized value of 100%. Other significant predictors included "time of purchase, I see myself as someone who is sometimes not shy and inhibited" (normalized importance of 52.2%) and "time of purchase, I see myself as someone who can be secure" (normalized importance of 44.3%). Conversely, predictors such as "time of purchase, I see myself as someone who is not easily stressed" and "time of purchase, I see myself as someone who does not worry a lot" had lower normalized importance values of 13% and 12.3%, respectively.

This implies that some consumers experience stress and worry at the time of purchase. FMCG marketers should therefore understand neuroticism concepts more deeply and develop strategies to offer alternatives that reduce stress and worry for consumers during their purchasing experience.

Implications for the Marketing Society of present system:
The system of present disclosure provides several valuable implications for the marketing industry, particularly in understanding and leveraging consumer neuroticism traits to enhance buying behaviour strategies for FMCG products:
1. Personalized Marketing Strategies:
o Tailored Messaging: Marketers can use insights into consumer neuroticism traits to craft personalized messages that address specific emotional needs. For instance, highlighting product features that alleviate stress or provide reassurance can be effective for consumers with high neuroticism.
o Targeted Promotions: By identifying traits such as "not getting nervous easily" as highly significant, marketers can develop targeted promotions that appeal to consumers seeking stability and confidence in their purchasing decisions.
2. Enhanced Consumer Segmentation:
o Behavioral Segmentation: The present disclosure provides more precise consumer segmentation based on neuroticism traits. This segmentation helps in creating marketing campaigns that resonate with different consumer groups, leading to more effective engagement and higher conversion rates.
o Product Positioning: Understanding the traits associated with higher importance, such as emotional stability, enables marketers to position their products in ways that align with these traits, thereby attracting consumers who value these characteristics.
3. Improved Product Development:
o Stress-Reducing Features: The high importance of traits related to emotional stability suggests a market opportunity for products designed to reduce stress and enhance well-being. Product development can focus on incorporating features that address these needs, such as calming ingredients or stress-relief functionalities.
o Consumer-Centric Design: Insights from neuroticism traits can guide the design of products that cater to the psychological needs of consumers, leading to innovations that better meet consumer expectations and preferences.
4. Effective Communication Channels:
o Optimized Channels: Knowing which traits are most influential can help in selecting the most effective communication channels. For example, consumers with high neuroticism may respond better to reassuring and empathetic communication through channels that emphasize trust and security.
o Message Framing: The study's results can guide how to frame marketing messages to appeal to consumers' emotional states, such as using calming and supportive language to engage those who are more prone to stress.
5. Strategic Customer Relationship Management (CRM):
o Personalized CRM: CRM systems can be enhanced to incorporate data on consumer neuroticism traits, allowing for more personalized interactions and support. This can lead to improved customer satisfaction and loyalty by addressing specific emotional and psychological needs.
o Behavioural Insights: Utilizing neuroticism traits in CRM strategies can help predict consumer behaviour more accurately, allowing for proactive management of customer relationships and more effective loyalty programs.
6. Ethical Considerations:
o Responsibility in Marketing: Marketers should be mindful of ethical considerations when leveraging consumer neuroticism traits. It's important to avoid manipulative practices and ensure that marketing strategies are designed to genuinely benefit and support consumers, rather than exploit their vulnerabilities.

The present disclosure provides the novel system for integrating psychological insights into marketing strategies to better understand and meet consumer needs. By applying these insights, marketers can create more effective, empathetic, and consumer-centric approaches that enhance both customer satisfaction and business outcomes.
, Claims:We Claim:

1. A system (100) for determining the buying behavior of products, the system (100) comprising:
ix. a neuroticism trait (101);
x. a neuroticism trait variable (102);
xi. an artificial neural network (103);
xii. a buying behavior module (104); and
wherein the buying behavior determination module determines the buying behaviors of the consumer (104) by applying an artificial neural network (ANN) (103) based on the value of each variable of the neuroticism traits (101) and the data associated with the consumer behaviors.

2. The system (100) as claimed in claim 1, wherein the neuroticism traits (101) comprise anxious, angry, depressed, self-conscious, impulsive, and vulnerable.

3. The system (100) as claimed in claim 1, wherein the neuroticism trait variables (102) labeled NEU-1 to NEU-9.

4. The system (100) as claimed in claim 1, wherein the artificial neural network (ANN) (103) is a multilayer perceptron (MLP) neural network, is to identify the neuroticism traits that influence the buying behavior of products by analyzing self-reported data.

5. The system (100) as claimed in claim 1, wherein the multilayer perceptron (MLP) neural network was automatically designed with three nodes in the hidden layer and four nodes in the output layer to represent the dependent variable, buying behavior.

6. The system (100) as claimed in claim 1, wherein the data were collected through a multi-stage random sampling method, followed by a survey to generate data for training and testing the ANN models.

7. A method for determining the buying behavior of product, the system (100) comprising:
vii. collecting and analyzing data associated with a set of consumer behaviors, wherein the consumer attributes neuroticism traits; and

viii. determining a product preference score for the consumer using an artificial neural network model based on the value of each neuroticism trait variable of the product; and
ix. the data associated with the set of consumer behaviors.



Dated this 28th October 2024

Documents