Evaluating the importance of different communication types in romantic tie prediction on social media

Abstract

The purpose of this paper is to evaluate which communication types on social media are most indicative for romantic tie prediction. In contrast to analyzing communication as a composite measure, we take a disaggregated approach by modeling separate measures for commenting, liking and tagging focused on an alter's status updates, photos, videos, check-ins, locations and links. To ensure that we have the best possible model we benchmark 8 classifiers using different data sampling techniques. The results indicate that we can predict romantic ties with very high accuracy. The top performing classification algorithm is adaboost with an accuracy of up to 97.89%, an AUC of up to 97.56%, a G-mean of up to 81.81%, and a F-measure of up to 81.45%. The top drivers of romantic ties were related to socio-demographic similarity and the frequency and recency of commenting, liking and tagging on photos, albums, videos and statuses. Previous research has largely focused on aggregate measures whereas this study focuses on disaggregate measures. Therefore, to the best of our knowledge, this study is the first to provide such an extensive analysis of romantic tie prediction on social media.

A Study of Machine-Learning Classifiers for Hypertension Based on Radial Pulse Wave

“…In each sequence, misclassified instances are given more weight for the next sequence while correctly classified instances are given lower weight. The final model is a linear combination of all the models created in the previous sequence [ 32 ]. In addition, GBDT has very few limitations and assumptions on the input data, so it is very flexible to deal with complex nonlinear relationships [ 33 ].…”

Predicting donation behavior: Acquisition modeling in the nonprofit sector using Facebook data

“…Since we are interested in classifying Facebook fans into donors and non-donors, we employ several prediction algorithms. We chose algorithms that have been proven to yield superior performance in analytical CRM and social media (e.g., [10,49] classifiers and is often used as a benchmark model in analytical CRM [18,21]. It estimates the conditional probability ( | ) by maximizing the likelihood function [40].…”

“…KNN makes no assumptions about the form of the mapping function other than that data points that are close are likely to have a similar outcome. It is equally a popular nonparametric method in social media applications [10].…”

“…AB iteratively gives more weight to misclassified observations [28], whereas XGB minimizes the gradient of the loss function in each iteration [29]. Both AB [8,9,10] and XGB [19] have proven to be successful in social media applications. Table 4 summarizes the settings of the hyperparameters of the different prediction algorithms.…”

Identifying Soccer Players on Facebook Through Predictive Analytics

“…For the entire collection of observations, 15.44% of the Facebook participants play soccer, and 84.66% do not. To cope with class imbalance, we oversampled the response variable (Bogaert et al 2016b). …”