Content uploaded by Md. Saddam Hossain
Author content
All content in this area was uploaded by Md. Saddam Hossain on Jul 15, 2021
Content may be subject to copyright.
Friends’ Influence Driven Users’ Value Change
Prediction from Social Media Usage
Md. Saddam Hossain Mukta1, Ahmed Shahriar Sakib2, Md. Adnanul Islam3,
Mohammed Eunus Ali4, Mohiuddin Ahmed5, and Mumshad Ahamed Rifat1
1United International University, Dhaka, Bangladesh
2American International University Bangladesh
3Military Institute of Science and Technology
4Bangladesh University of Engineering and Technology
5Edith Cowan University, Australia
Abstract. Basic human values represent a set of values such as security,
independence, success, kindness, and pleasure, which we deem important
to our lives. The value priority of a person may change over time due
to different factors such as life experiences, influence, social structure
and technology. In this study, we show that we can predict the value
change of a person by considering both the influence of her friends
and her social media usage. This is the first work in the literature
that relates the influence of social media friends on the human value
dynamics of a user. We propose a Bounded Confidence Model (BCM)
based value dynamics model from 275 different ego networks in Facebook
that predicts how social influence may persuade a person to change
her value over time. Then, to predict better, we use a particle swarm
optimization based hyperparameter tuning technique. We observe that
these optimized hyperparameters produce more accurate future value
score. We also run our approach with different machine learning based
methods and find support vector regressor (SVR) outperforms other
regressor models. By using SVR with the best hyperparameters of BCM
model, we find the lowest Mean Squared Error (MSE) score as 0.00347.
Keywords: Values ·Facebook Friends ·Influence ·BCM ·
Hyperparameters ·PSO.
1 Introduction
In recent times, Social Networking Sites (SNS) have become a major platform of
communications among users on the web. These SNS data provide a wide range
of opportunities to identify cognitive and psychological attributes such as basic
human values (aka values) [8], and personality [16] of users. Values represent
one’s attitudes, opinions, thoughts, and goals in life. Values of an individual
might change over time due to the influence of her friends [10,14]. In this paper,
we predict users’ value dynamics (change of value scores over time) based on her
2 Mukta et al.
social media usage (e.g., statuses and comments) and her friends’ influence in
an egocentric network such as Facebook6.
Values are essentially a set of criteria such as security, self-enhancement,
etc., that influence individuals to take different actions. Chen et al. [8] predict
five higher-level values from user word usages in Reddit. In another study [19],
authors predict users’ values from generated (i.e., statuses) and supported (i.e.,
likes and shares) contents in Facebook. However, these approaches largely fail
to capture the change users’ value priorities (over time) from the social network
usage. Several socio-psychological studies [9,14] show that values of a user might
be reshaped by the influence of other members in the same interest group. To
the best of our knowledge, no study has been conducted to identify the change
of value scores of users that considers both the influence of her friends and
her social network interactions. Identifying the change of values of an individual
from friends’ influence has a number of applications such as identifying university
course major or career path shifting trends, prediction of customers’ purchasing
behavior, transforming customers’ product selection preference or marketing
policies, and understanding transition of economics and business.
Based on the above observation, we propose a technique to identify the
changes of a user’s value from her social network interactions with friends by
using Bounded Confidence Model (BCM) [22,11,17]. Users’ value might change
due to other factors; however, for friends’ influence calculation we use our modified
version of popular BCM model. We observe the users’ change of values in a time
span of 10 years. Motivated by the work [8], first we collect data of 275 user
networks from Facebook by using a google survey form. We mainly identify
the ego networks with their close friends, i.e., N≤5[1]. Since users’ change
of values are observed in terms of time intervals, we separate users’ Facebook
statuses, comments, and shares according to a time span of six months [21].
Then, we generate value scores from users’ each six months interval Facebook
statuses, comments, and shares by using IBM Watson Personality insight API 7.
From the computed value scores of users’ writing content, we compute the
hyperparameters such as convergent factor, µand threshold, and σ, for our
modified BCM model (see Equation 1). Then, we use particle swarm optimization
(PSO) [4] method for finding the best hyperparameter configuration. Finally,
we use these optimized hyperparameters of the BCM model to predict next
value score by using support vector regressor (SVR) model [3]. In summary, our
contributions in this paper can be highlighted as follows:
–We are the first to propose a value change identification technique that
predicts user value change dynamics by considering the influence of friends
and her social media interaction.
6Facebook is an example of an egocentric network [2] because the network provides
interaction capability only among the friends while preventing any interactions from
the external users to this network.
7https://cloud.ibm.com/catalog/services/personality-insightsabout
Value Changes by Friends’ Influence 3
–We modify the BCM model and capture the friends influence in an ego
centric network.
–We develop a PSO based best hyperparameter selection method that predicts
user’s future value score with less MSE score.
2 Preliminaries and Related Work
2.1 Values
Basic human values define the goal, belief and behavior of an individual.
Schwartz et al. [24] categorize the value dimension into five higher-level values.
Openness-to-change mainly refers to the ability to “think outside of the box”
which consists of two broad personal values: self-direction and stimulation [23].
Self-transcendence satisfies the motivational goal for preservation and
enhancement of the welfare of people in the society [23]. Self-enhancement
represents a person’s interest to be socially recognized and attraction for control
over other humans and resources in the society. Conservation emphasizes order,
self-restriction, preservation of the past, and resistance to change. Hedonism
basically means pleasure or sensuous gratification for oneself [23].
2.2 Modified Bounded Confidence Model (BCM)
BCM is a popular opinion dynamics model to determine the influence of a
network of people over an individual. Motivated by the BCM devised by Deffuant
et al. [12], we revise the model (see Equation 1) and introduce users’ value change
model. Considering the distance of the corresponding values between two users
(i.e., an ego and each alter) is less than a given threshold σ, the updated value
of each of the users can be computed by using the following equation:
BH V t+1
i=BH V t
i+µego
ΣN
j=1(BHV tj−BH V ti)
NΘ(σego−| ΣN
j=1(BHV tj−BH V ti)
N|)
(1)
Here, BHViis the value score of the user i,jrepresents her friends, µego is a
convergence factor, σego is the threshold within which the users interact or adapt
with each other, and Θ() is a Heaviside’s theta function 8. According to Dunbar
number, friends can influence who fall in the intimate sphere where N≤5[2].
2.3 Values in Social Media
Chen et al. [8] identify values from Reddit, an online news sharing community.
The authors identify five higher-level values from user’s pattern of word using
in social media. They predict the value scores by using linear regression. Boyd
et al. [7] identify values from statuses of 767 Facebook users. They identify
values with a data driven approach. Mukta et al. [19] identify values from both
8https://mathworld.wolfram.com/HeavisideStepFunction.html
4 Mukta et al.
user generated and supported contents in Facebook. Mukta et al. [21] predict
the temporal change of values of Facebook users by using hybrid LSTM model.
In this paper, we devise a novel technique to identify the value change of an
individual by the social influence of her friends in Facebook.
2.4 Value Changing Influence Models
The value of a user might be changed due to different offline behaviors such as
life experiences, life events, technological change, social structure, and life style
of others [15] [10]. In previous studies [21,20], we find that value change of an
individual reflects in their social media usage behavior, where authors capture
the opinion change over time from social media. Quattrociocchi et al. [22] show
that inner dynamics of information systems, i.e., TV, newspaper, social network
platforms, - play a vital role on the evolution of the public network. In this paper,
we compute value scores in different time intervals and analyze the interactions
among users. Then, we measure how the value of one person may influence the
change of value of others through social media interactions.
3 Methodology
In this section, we first describe the process of value change modeling by the
influence of close friends using BCM model. Figure 1 shows the complete pipeline
of our value change modeling and its hyperparameter optimization process. We
discuss our methodology in the subsequent sections.
Fig. 1: Methodology of value change modeling and BCM hyperparameter optimization.
3.1 Data Collection
We randomly select a total of 275 (motivated by the study of Golbeck [16])
different Facebook ego networks where each user holds an ego network and his
Value Changes by Friends’ Influence 5
Facebook friends are alters. Initially, we randomly invite a total of 320 Facebook
friends to share their ego networks, but only a total of 275 users show their
interest to share their data. After selection, we extract the list of users; then
collect every comment of each user. Users’ collect their data from their own
profile 9and download her data 10 in different time intervals with JSON format.
Later, they share their data with us through the google form.
We extract public profile data like statuses, posts and shares that support
comments and likes. Then, we divide the data into temporal segments. Each
year is divided by 2 segments, each segment contains data of 6 months. In our
experiment, we have Facebook posts of maximum 10 years for a single user,
which we can divide into 20 segments. We select the users based on the number
of their daily public interactions. Table 1 shows the statistics of our dataset.
Table 1: Statistics of our dataset
Attributes Values
Number of ego Networks 275
Number of total Comments 75,625
Number of Maximum Comments of a User 237
Number of Minimum Comments of a User 25
Total Time Duration (years) 10
Maximum Time Span for a user (years) 10
Minimum Time Span for a user (years) 7
3.2 Influence Modeling for Value Change
Next, we propose a new technique to investigate the change of values over time.
To do so, we apply the BCM model [12], where we use two hyperparameters:
convergence factor (µ)and threshold (σ)of value difference. To optimize these
hyperparameters, we use a machine learning based approach to determine the
optimum values (i.e., solution) by using PSO algorithm (Figure 1). We actually
tune the hyperparameters of the regression models. From these optimized
hyperparameters, we find the threshold (σ)value for predicting the next value
score through using SVR.
4 Experimental Evaluation
4.1 Ego Network Configuration
We consider the adjustment of the value score of an ego (ui) by interacting
all users uj. We collect interaction data such as likes, comments, and sharing
of object between an ego u, and their friends, ujwhere j≤5according to the
Dunbar number [2]. We are interested in showing that when a number of users
ujinfluence to value score of a single user (e.g. an ego), her value score converges
to a single unified score [6]. Hence, in this study, we assume that users may shift
their value scores by the influence of close friends who fall in the sphere where
N≤5[2].
9https://www.facebook.com/settings?tab=your_facebook_information
10 https://www.facebook.com/dyi/?referrer=yfi_settings
6 Mukta et al.
Fig. 2: Group wise influence based value change from time t1(Fig a) to t2(Fig b)
Similar to a group discussion, we select a primary user to detect his/her value
change from the influence of other people of the group. Figure 2(a) shows that
independent value scores for a total of 5 Facebook users. When they interact with
each other in a group, value score for a user might be changed by the influence of
other alters, i.e., Facebook friends. Figure 2(b) shows the final score for a value
dimension after being influenced by the members of the group. We predict this
change of user’s value by using our modified BCM model. The model has different
hyperparameters that we optimize by using Particle Swarm Optimization (PSO)
method as discussed next.
4.2 Hyperparameter Estimation
During hyperparamer estimation, for brevity and simplicity we only consider
hedonism value dimension among all top-level 5 value dimensions. We collect
users’ pleasure seeking behavior, i.e., hedonism, by easily extracting their digital
footprints (i.e., restaurant/movie check-ins, purchasing a gadget, etc.) from
Facebook. We could also collect users’ food related data from their check-ins of
different restaurants. The convergence factor, µindicates the momentum term
of the influence dynamics. In our study, we consider µ= 0.4 following the study
of [22]. However, to obtain the appropriate value of the threshold (σ) for the
BCM model, we build a regression model. The regression model actually predicts
the σwhich minimizes the error to the least. To build the regression model, we
consider four features for the training instances - i) value of user, ui, at time
t, ii) value of user’s friend, uj, at time t, iii) value of user, ui, at time t+1,
and iv) convergence Factor, µ. We perform several regression models with these
features by a 10-fold cross-validation with 10 iterations. We use the following
Value Changes by Friends’ Influence 7
regressors: SVR, Gaussian Process Regressor (GPR), ElasticNet, BayesianRidge,
and MLPRegressor (MLPR) 11.
For hyperparameter tuning, we apply PSO 12 by using Optunity Library 13.
PSO [4] is a typical algorithm of the swarm intelligence family. The algorithm
is a population-based meta-heuristic optimization technique which initializes a
number of individual search ‘particles’, each representing a possible solution.
This population of particles change their positions by an evolutionary process.
Each of these particles is in movement with a velocity allowing them to update
their position over the iterations to find the global minimum. ParticleSwarm 14
has 5 parameters that can be configured: num_particles, num_generations,
φ1,φ2,and max_speed. In our experiment, we use num_particles = 10 ,
num_generations = 15 ,max_speed = None, φ1= 1.5 , φ2= 2.0 to initialize
ParticleSwarm 15 solver in Optunity. In our case, we consider one set of
hyperparameter configuration as a single particle.
We split the dataset into training, and test data by 70% (a total of 193
instances) and 30% (a total of 82 instances), respectively. Then, we define our
objective function to minimize the cost of the model which in this case is,
Mean Squared Error (MSE). We initialize different box constrained configuration
sets of hyperparameters for different regressors. Each particle represents a
configuration for hyperparameters of the machine learning model. All of the
particles have MSE, i.e., fitness values, which are evaluated by the cost function
to be minimized. The particles move through the problem space by following
the current optimum particles. Table 2 shows the optimized hyperparameter
configuration values for different regressors.
Table 2: Hyperparameters configuration for different regressors
Regression Model Hyperparameter Configuration
SVR
kernel -RBF, gamma- [0, 50] , C-[1, 100] ;
kernel -linear,C-[1, 100] ;
kernel -poly, degree- [2, 5] , C-[1000, 20000], coef0-[0,1]
Gaussian Processes normalize_y= [True,False], alpha = [1e-10 - 1e-2]
ElasticNet alpha - [0, 1.0] , l1_ratio - [0, 1.0] , tol - [1e-4, 0.01]
BayesianRidge alpha_1 - [1e-6, 0.01], alpha_2 - [1e-6, 0.01],
lambda_1 - [1e-6, 0.01], lambda_2 = [1e-6, 0.01], tol -[1e-4, 0.01]
MLPRegressor hidden_layer_sizes - [( 50, 50, 50), ( 50, 100, 50 ), ( 100, )] ,
activation - [’tanh’, ’relu’], alpha - [1e-4, 0.01]
Since Optunity can optimize conditional search spaces, we set different
hyparameters for SVR based on the kernel (e.g., Radial Basis Function (RBF),
linear, and polynomial. Table 2 shows the search space for SVR hyperparameters.
Moreover, Table 3 shows the performance of the regressors over our test dataset.
11 SVR: https://bit.ly/2OOBDZa, GPR: https://bit.ly/2OWxAKy, ElasticNet:
https://bit.ly/3pDuh7M, BayesianRidge: https://bit.ly/3qJoyys, MLPR:
https://bit.ly/3qEivez
12 https://bit.ly/3ucqcLf
13 https://homes.esat.kuleuven.be/ claesenm/optunity/
14 https://bit.ly/37zcGYb
15 https://bit.ly/3aBdSw3
8 Mukta et al.
Table 3: Strength (Low RMSE) of the regression model
Regression Algorithm MSE
SVR 0.00334
Gaussian Process Regressor 0.07138
MLPRegressor 0.08912
ElasticNet 0.05312
BayesianRidge 0.01183
We find SVR regressor shows the best average performance (MSE-0.00337) to
predict users’ threshold value on the BCM model. Algorithm 1 presents the
process of our PSO based hyperparameter optimization method.
Algorithm 1: PSO_based_BCM_Hyperpar_Optimization
initialize:
uias value of a user
uj. . . n as values of group users
ui+tas value of a user in
time t
µas convergence factor ( 0.4 )
σas Threshold Value
Proc.: Dataset_Prep_BCM( )
1: Calculate the threshold value via BCM
Model
2: Set predictor variables (X) : ui,uj.. . n ,
ui+t,µ
3: Set Dependent variable : σ
4: Divide the Data set by 70% and 30%
as D1, and D2, Respectively
Proc.: Model_Training( )
5: initialize D1 as train dataset
6: initialize D2 as test dataset
7: Run ML models to predict σ
Proc.: Hyperparam_Optim( )
PSO: Initialize
num_particles=10
num_generations=15
max_speed=None
φ1=1.5
φ2=2.0
8: Find the optimum σfor BCM model
using PSO
9: Save ML model as M1
10: Calculate the ui+tusing M1
11: Calculate MSE on D2 dataset with
respect to M1
5 Results and Discussion
In this experiment, we take Facebook interactions (i.e., statuses, comments,
shares, and likes) of 275 users. By using PSO, we fine tune the parameters of the
estimators to get the optimized threshold. With these parameters, we predict
the best threshold, σto predict the accurate value scores by using SVR.
Next, Figure 3 presents the amount of loss when compute the σby using
different models. Among these models, SVR shows the lowest loss to predict the
hyperparameters. Use this threshold, σvalue, we predict the best final values
score of a user influenced by her close friends. Figure 4 shows the actual future
value scores and predicted value scores using SVR for different users.
Berndt [5] describes that friendships have influence on user’s attitude and
behavior. For example, adolescents whose friends drink beer at parties likely to
Value Changes by Friends’ Influence 9
Fig. 3: Comparison of loss among the
regression models.
Fig. 4: Comparison between users’
actual and predicted value scores.
start drinking. In contrast, user’s value may influence negatively. For example,
adolescents often have conflicts with others which might propagate among others.
Epstein [13] and Hartup [18] describe that friends influence each other in different
behaviors, including aspirations, achievements, values and attitudes. Our study
also shows that friends’ influence can persuade to change one’s value which can
even be predicable from the social media usage.
6 Conclusion
In this paper, we have extracted 275 different ego networks from a Facebook.
Then, we have identified intimate friends (N≤5) for each of the ego networks.
Later, we have segmented users’ interaction in a time frame of 6 months.
Then, we have computed users’ value scores from their Facebook interactions
by using IBM personality insight API. Based on the users’ value scores, we have
proposed a value dynamic technique based on modified BCM influence model.
During modeling, we have also proposed a PSO based hyperparameter estimation
technique. Our model have showed an outstanding performance (i.e., lower MSE)
in predicting change of users’ value from their Facebook interactions.
References
1. Arnaboldi, V., Conti, M., Passarella, A., Dunbar, R.: Dynamics of personal social
relationships in online social networks: a study on twitter. In: Proceedings of the
first ACM conference on Online social networks. pp. 15–26. ACM (2013)
2. Arnaboldi, V., Guazzini, A., Passarella, A.: Egocentric online social networks:
Analysis of key features and prediction of tie strength in facebook. Computer
Communications 36(10-11), 1130–1144 (2013)
3. Awad, M., Khanna, R.: Support vector regression. In: Efficient learning machines,
pp. 67–80. Springer (2015)
4. Bansal, J.C.: Particle swarm optimization. In: Evolutionary and swarm intelligence
algorithms, pp. 11–23. Springer (2019)
10 Mukta et al.
5. Berndt, T.J.: Friendship and friends’ influence in adolescence. Current directions
in psychological science 1(5), 156–159 (1992)
6. Boudec, J.L., McDonald, D., Mundinger, J.: A generic mean field convergence
result for systems of interacting objects. In: QEST. pp. 3–18 (2007)
7. Boyd, R.L., Wilson, S.R., Pennebaker, J.W., Kosinski, M., Stillwell, D.J., Mihalcea,
R.: Values in words: Using language to evaluate and understand personal values.
In: ICWSM (2015)
8. Chen, J., Hsieh, G., Mahmud, J.U., Nichols, J.: Understanding individuals’
personal values from social media word use. In: CSCW. pp. 405–414. ACM (2014)
9. Crandall, D., Cosley, D., Huttenlocher, D., Kleinberg, J., Suri, S.: Feedback effects
between similarity and social influence in online communities. In: SIGKDD. pp.
160–168. ACM (2008)
10. Danis, W.M., Liu, L.A., Vacek, J.: Values and upward influence strategies in
transition: Evidence from the czech republic. JCCP 42(2), 288–306 (2011)
11. Deffuant, G., Amblard, F., Weisbuch, G.: Modelling group opinion shift to extreme:
the smooth bounded confidence model. arXiv preprint cond-mat/0410199 (2004)
12. Deffuant, G., Neau, D., Amblard, F., Weisbuch, G.: Mixing beliefs among
interacting agents. Advances in Complex Systems 3(01n04), 87–98 (2000)
13. Epstein, J.L.: The influence of friends on achievement and affective outcomes. In:
Friends in school, pp. 177–200. Elsevier (1983)
14. Friedkin, N.E., Johnsen, E.C.: Social influence and opinions. Journal of
Mathematical Sociology 15(3-4), 193–206 (1990)
15. Friedkin, N.E., Johnsen, E.C.: Social influence networks and opinion change.
Advances in group processes pp. 1–29 (1999)
16. Golbeck, J., Robles, C., Edmondson, M., Turner, K.: Predicting personality from
twitter. In: PASSAT. pp. 149–156. IEEE (2011)
17. Gómez-Serrano, J., Graham, C., Le Boudec, J.Y.: The bounded confidence model
of opinion dynamics. Math. Models Methods Appl. Sci 22(02), 1150007 (2012)
18. Hartup, W.W.: Adolescents and their friends. New directions for child and
adolescent development 1993(60), 3–22 (1993)
19. Mukta, M.S.H., Ali, M.E., Mahmud, J.: User generated vs. supported contents:
Which one can better predict basic human values? In: International Conference on
Social Informatics. pp. 454–470. Springer (2016)
20. Mukta, M.S.H., Ali, M.E., Mahmud, J.: Identifying and predicting temporal change
of basic human values from social network usage. In: Proceedings of the 2017
IEEE/ACM International Conference on Advances in Social Networks Analysis
and Mining 2017. pp. 619–620 (2017)
21. Mukta, M.S.H., Ali, M.E., Mahmud, J.: Temporal modeling of basic human values
from social network usage. JASIST 70(2), 151–163 (2019)
22. Quattrociocchi, W., Caldarelli, G., Scala, A.: Opinion dynamics on interacting
networks: media competition and social influence. Scientific reports 4(2014)
23. Schwartz, S.H.: Universals in the content and structure of values: Theoretical
advances and empirical tests in 20 countries. Advances in experimental social
psychology 25(1), 1–65 (1992)
24. Schwartz, S.H.: A proposal for measuring value orientations across nations.
Questionnaire Package of the ESS pp. 259–290 (2003)
