Gender Prediction for Instagram User Profiling using Deep Learning
Adri Priadana, Muhammad Rifqi Maarif, Muhammad Habibi · 2020 International Conference on Decision Aid Sciences and Application (DASA) · 2020
Instagram creates new opportunities for small businesses to expand their markets. It can be used by entrepreneurs to reach potential customers and tell them about their products. Hence, knowing the user's demographics on Instagram is essential to convert those users into potential buyers. The demographics of social media users, such as gender, are vital for personalized advertising targeting. Based on the previous study, most of them studied to predict gender based on text analysis. This study aims to implement a deep learning method called Convolutional Neural Network (CNN) to predict gender on Instagram based on an Instagram profile image. Deep learning is a widely known technique to extract hidden patterns of a specific image. Thus, it can be useful for detecting gender based on the Instagram user's profile images. According to performance analysis, the prediction of gender on Instagram based on profile pictures using CNN resulted in an accuracy value of 70.11%. Compared with research related to gender prediction based on text analysis in previous studies, this study's accuracy results are not better than the four different studies in the state-of-the-art. However, this study's accuracy results are better than the two other studies in the state-of-the-art. Furthermore, this study proves that gender prediction based on image analysis using the CNN method can be performed exceptionally well, especially image analysis based on the image profile of Instagram users.