Keywords

1 Introduction

The scientific study of the human mind and its functions, particularly those that influence behavior in a certain situation, the mental traits or attitude of a person or group are coined as Psychology. It even includes the investigation of both conscious and unconscious occurrences, as well as feelings and thoughts; a vast academic field with several possibilities. A professional practitioner or researcher who imbibes these disciplines is called a psychologist. Moreover, they engage themselves in exploring the physiological processes that underlie cognitive functions and behaviors. Incoming thoughts in our mind that are unstoppable with due reasons have a significant impact on our subconscious mind resulting in the tendency that an individual is passing through. They are specialized in either social, behavioral, or cognitive sciences background. Our research discusses the behavioral aspect of psychology, often known as “Behavioral Psychology”. Some psychologists try to figure out what role mental functions play in personal and societal behavior

Behavioral psychology has an adverse impact on the personality traits of a person. Based upon the personality traits identified by the five-factor model [1], the most important predictors of health behavior are conscientiousness and neuroticism [2, 3]. Individuals with greater levels of conscientiousness, for example, are more likely to exercise [4], adhere better to medication [5], likely to obtain preventive cancer screenings [2]. Figure 1 shows a five-factor model. Those with higher neuroticism, undergo mental trauma since it is characterized by emotional deregulation or high emotional reactivity.

We provide a psychological-behavioral model and an experiment based on it in this paper. We also go through the behavioral modification techniques that are based on this paradigm. Machine learning classification algorithms mainly Decision Tree Classifier, Random Forest Classifier [6], Gaussian Naive Bayes classifier, and Support Vector Machine [7] are applied on obtained preprocessed data to gain meaningful insights based on the target variable and feature set considered. Results obtained after training these models on the preprocessed data clearly state the psychology of an individual and behavior as a whole. The next chapters of this research brief about the related work done in this domain, followed by Methodologies, Results and discussions. Conclusion and Future Scope end our research.

2 Related Work

Yang Huanhuan [8] aim at understanding the behavioral habits of college students and cultivate good behavioral habits in them. The paper raises the issue of how nowadays people not just aim for food and clothes but also desire for a happy life and this is where positive psychology comes into the picture to understand the developmental behavior of the person with respect to its societal presence and modernization. This the paper thoroughly talks about behavior and psychology as a good developmental base for college students. While talking about the importance of developing good developing habits in college students, this paper talks about how moral habits can help in personal accomplishment for an individual college student and it directly or indirectly helps national and social development. While discussing problems existing in college students it discusses how lack of moral behavior is present in those students who tend to be influenced by the internet which results in a bad moral attitude of an individual. It also raises the issue of students facing problems in learning behavior where it talks about how weak self-control can develop bad habits which develop bad life behavior which consists of improper scheduling, lack of moral values, becoming constantly angry and jealous of anything. The main reasons for this highlighted by the paper areas lack proper upbringing, school neglecting students’ behavior. For improvement in positive psychology in college students, the paper talks about how families and schools working together can help in the proper behavior development of the students.

Gomez Lopez et al. [9] discuss how love and relationships are either playing a key role in developing well-being or are becoming a negative outcome. According to the findings, romantic relationships can provide a source of happiness for teenagers and emerging adults. Study shows that relationships can improve social integration and can create a positive self-concept. The studies included in the qualitative synthesis were around 112 and SPSS from IBM was used to carry out coding and process results. From the complete study of more than 100 researchers previously done a finding which came to existence is that a relationship which is beneficial for well-being, have high-quality levels, partners developing their personal and shared goals maintain a secured attachment over some time. Personal well-being is the only key to happy and satisfied relationships of a person’s life and people with disappointing or broken relationships state to face failure or depressed phase of their life or ending up they blame their previous relationships.

Astatke et al. [10] aimed to identify the correlation of emotional intelligence, academic help to seek behavior, and psychological phenomena among the 283 students of first-year regular diploma. For correlation, SPSS was used and while creating questionnaires 3 major parameters of emotional intelligence, help-seeking behavior, and attitudes towards professionals were checked. Data gathering permission from the college authorities was taken and researchers administered the questionnaire before permitting to gather data. The results showed that emotional intelligence had a significant positive correlation with students having academic achievement, it also came to a conclusion that students’ emotional intelligence increases with their better academic performance.

W J Chopik [11] explored findings done by two studies about relationships in which the first involves more than 270000 people in nearly 100 countries. It is identified that family and friend relationships are associated with a healthy and happy lifestyle. From the second study done on 7500 aged people in us, it was identified that not just having friends but the quality of friends also mattered and it has played a major role in creating a behavioral personality. The findings showed that for aged people across the time prioritize friendship more than family relationship, the people who claim to be self-prioritized are the one who has been part of temporal friendships or had lost faith in the family, people generally tend to stay with friends and family at most, and they are somewhat sharing a powerful friendship with their family members as well. Concluding the result, we can say that friendship and family are not distinct at all.

Valkenburg, Patti [12] investigates how social media can influence the self-effects of the message creator(involuntarily). Self-effects can be defined as effects on cognition (knowledge or beliefs), emotions, and attitudes by some external factors. This can be seen from an experimental study where an individual is told to write a blog about something, this creation of messages not only affects the qualities (such as cognition, emotions) of readers but, it can also affect the qualities) and traits of the message sender/creator. The author concludes by saying that self-effects are stronger in online settings rather than in offline settings.

3 Methodology

Fig. 1.
figure 1

Five-factor model

Full size image

3.1 Sample and Data Collection

The study population is made for all over India. The people were aged between 15 and 71 where we classified them into three age groups Youth, Adults, Millennials. The majority of the respondent were between the age group from 19 to 24 falling in classes of adult and youth. While rest response was gathered from Company employees and Family members which fall in a range between 40 and 60. A total of 19 questions were asked to 200 respondents out of valid data set of 175 responses were used for data analysis.

In this study, the majority of questionnaires were Binary be- tween, yes and no, couple of them were questions based upon single choice with multiple options. And the last part consists of the multiple-choice question where the respondent is supposed to choose all the possible options on a given scale which is applied to them. Questions majorly include personality and behavioral type too. Figure 2. includes questions asked in the survey.

Fig. 2.
figure 2

Survey taxonomy

Full size image

3.2 Data Analytics

The main propaganda of this research is to analyze the responses obtained from the known youth of the university and professionals(faculties). The survey was conducted from May 4, 2021, till 10th May 2021. A sufficient number of responses were obtained with satisfactory answers and no manipulations. The purpose of approaching known personnel was thereby met, and we were pretty sure that what type of answers he/she might give to the respective questionnaire.

The complete survey taxonomy was divided into 2 categories as:

  • Social Media Influence of the subject and its adverse impact on behavioral psychology, his decision over social media validation as well.

  • General choice-based questions answering its psychology. Figure 3. is the process flow we followed. From the data received, 17.91% were youth, 72.83% as adults, and 9.26% as millennial.

Fig. 3.
figure 3

Architectural block diagram

Full size image

3.3 Technology Stack

Following is the technology which we worked on:

  • Pre-processing – Before entering the data into our model we need to conduct some pre-processing over data with the help of cross tab, label encoding, and one hot-coding. First, over the data, we applied the cross tab to point out the relations among multiple conditions and determine the important factor which is more prominent in our data set. Secondly, we applied the label encoding to the columns of an object type to int which will be further required for the Machine algorithm model. Lastly, we applied one hot coding for the specific columns.

  • Gaussian Bayes – Gaussian Bayesian classification is used for normal distribution for continuous valued features with no co variance in them, with help of Gaussian classification we are successfully able to identify relationships among various feature values. Label encoding for binary vales and one-hot encoding for categorized value is done, after encoding a target variable is selected and based upon its relationship with other variables features selection is done, after selection train, test, split method is implemented on the data set and Gaussian classifier is called for fit in training and testing data. Using scikit-learn metrics function accuracy is identified.

  • Random Forest regression – We utilized Random Forest Regression, a machine learning model that can be used for both regression and classification, to assess the significance of characteristics within the subset and apply it to the data set. We used random forest regression to predict the BMI value in relation to the specific field, and we got a lot of different outcomes with an accuracy of above 95%.

  • Decision Tree – A very specific probability tree that allows you to create a decision about the process flow given. After applying label encoding and one hot-encoding we applied the decision tree on our data set where we selected the infection lingered as our target field and classified successfully with an accuracy of 86%.

  • Adaboost Algorithm – A primitive classification technique also known as adaptive boosting algorithm is an Ensemble method in machine-learning which is gradually used to improve the model’s accuracy.

  • XGBoost Algorithm – Another preemptive ensemble method is XGBoost. It is a decision-tree based algorithm which is designed for speed and performance by implementing gradient-boosted decision trees.

4 Results

4.1 Case Studies

Case-1: - Responses received from the youth state that, we belong to such generation where social validation matters a lot. Table 1. Consists of the results and the feature set as well the target variable considered. On the other hand; adults do feel the same as youths when social validation is considered.

Table 1. Behavior of youth, adults and millennials regarding social validation.
Full size table

Case-2: - In this we considered, 4 questions to predict the behavior of members of respective age category i.e., youth, adults, and millennials. “Verdict to be hanged”, “Whom will you choose APJ Abdul Kalam or Donald Trump”, “What do you prefer as your ultimate choice from Success, wealth, love, peace” and “Argumentative situation-based decision” are those questions used. Our model proposes that people who follow APJ Abdul Kalam; Are in favor of that “verdict should be hanged” which ultimately shows a sign of positive correlation and people with this mindset have an inclination towards good psychology. On the contrary, people who follow Dr. APJ Abdul Kalam sir as their role model; answered the argumentative question as “Will Remain Silent; accept and will try to evaluate our mistake” which symbolizes the characteristics of that ideal person. Table 2 showcases the results generated by classification algorithms for this case.

Table 2. People responding to such situations by the influence of role model
Full size table

Case 3: - Another stack of questionnaires considered is “Verdict to be hanged or released”, “Argumentative situation- the based question”, “Priority” and your nature “Are you a Giver, Matcher or Taker” [13] as feature set and “Do you feel satisfied after hurting someone” as the target variable. Table 3 includes the results of classification machine learning algorithms on these attributes. Our model can now predict the solution an individual will propose to the target variable based question on the results obtained for the feature set’s questionnaire.

Table 3. Individual’s aggressive behavior based on a scenario
Full size table

4.2 Findings

Extensive features of WhatsApp make it most in-demand social-media platform. Table 4. includes presence of respondents and their personal preferences while social media is considered. Instagram followed by snapchat is what our results state. Table 5 includes insights gathered by concatenation of choices gathered from multiple questions.

Table 4. Social media presence of people responded
Full size table
Table 5. Meaningful insights from the obtained data
Full size table

4.3 Visualiations

Visualizations are a way of showing the data into a pictorial form which is the best way to visualize the data and allows us to get clear insight withing the data. It allows us to explicitly interact the data and see data through the visualization tool named as Tableau. This helps us to create an interactive dashboard with the visualization created on it. The visualization given below are created on the survey of the data which we generated.

Fig. 4.
figure 4

Classification of choices with respect to age

Full size image

In Fig. 4. we plotted a tree map depicting the relation of priority-based response of “What will you choose from Success, Wealth, Love, Peace” and all the subjects with age categories they fall into. To our surprise, we found that; Age group 20–22 want peace in their life. Age group between 15–19 prefer the rest over peace.

Fig. 5.
figure 5

Influence of ideal person on issue-based scenarios

Full size image

In Fig. 5. we got the graph between the ideal person i.e. “The person you follow”; the orange block is indicated with the responses of the question “Verdict to be hanged” and blue indicates whether to “Give him a chance” where we were able to justify that people chose such responses that truly resemble the characteristics of the role model they follow. Almost all who chose Donald Trump as an ideal person, responded that Verdict should be hanged which depicts aggressive mindset of an individual.

5 Implication and Limitation

After Pandemic and change in work culture, majority of the crowd is finding the change in their behavior and thinking patter which we believe is a major concern of the society. People’s mindsets have changed even their pattern of thinking after moulding to the latter lifestyle.

However, this study has many limitations. First unique finding of this study is that it has identified the relation between the behavior and thinking patter over a specific group of crowd and age specifically. Although there are some limitations in this research, though the results we obtained through the survey are specific and need to be judged when we meet these individuals in personal through which we can obtain more precise responses and can clearly see their traits of personality and behavior over the year. Second the majority of the crowd volunteered belongs to age group of 19 to 24 while we received very few responses for the category of Youth’s and Millennials, hence we aren’t that fortunate to talk upon these age groups.

6 Conclusion and Future Scope

This extensive study of juveniles, adults and millennials thereby claimed that irrespective of age, evaluating the psychological behavior is not a cakewalk and consequential as a whole. The survey of 16 questions proposed meaningful insights and subsequent relation of questions, showcased high quality preciseness. The received responses of the subjects indirectly state the personality trait; hence following the five-factor model. Results are obtained using machine learning algorithms specifically classification algorithms such as Decision Tree Classifier, Gaussian Naive Bayes Algorithm, Random Forest Classifier and Support Vector Machine. We achieved a considerable accuracy of 92.30% i.e., our model can now predict the choice of 92 people in 100 individuals whether they will remain silent or react forcefully on an argumentative situation. Gaussian Naive Bayes classifier gives commendable results.

Since the entire data obtained is of textual type, analyzing the sentiments behind the responses will allow us to understand the psychology of the individual by using the Natural language processing-based algorithms and we thereby aim to extend our research to that extent which eventually will provide more accurate and precise results.