Academia.eduAcademia.edu

Figure from:Predicting Movie Genre Preferences from Personality and Values of Social Media Users

See full PDFdownloadDownload figure
Table 1: Personality based classification of movie genres

Table 1 Personality based classification of movie genres

Related Figures (3)

ferent classifiers from personality and values to predict users’ movie genre preferences that are described in previ- ous sections. Since we train different classifiers that produce weights, we compute weighted linear ensemble score using the weights in Table 3. Finally, we build our weighted linear ensemble model using the weights generated from another dataset, so that our models do not get over-fitted. Table 4 presents the movie genre preference classification result by using the ensemble of personality and values. We observe that the average AUC of our classifier is 63.4%, and the baseline accuracy is 56.2%. We also observe that our clas- sifiers largely outperform the random baseline. We also find that MAE of our model is 0.18. We find that our ensemble of classifiers achieves higher accuracy than the independent personality and value based classifiers.
ferent classifiers from personality and values to predict users’ movie genre preferences that are described in previ- ous sections. Since we train different classifiers that produce weights, we compute weighted linear ensemble score using the weights in Table 3. Finally, we build our weighted linear ensemble model using the weights generated from another dataset, so that our models do not get over-fitted. Table 4 presents the movie genre preference classification result by using the ensemble of personality and values. We observe that the average AUC of our classifier is 63.4%, and the baseline accuracy is 56.2%. We also observe that our clas- sifiers largely outperform the random baseline. We also find that MAE of our model is 0.18. We find that our ensemble of classifiers achieves higher accuracy than the independent personality and value based classifiers.