Clustering Arabic Tweets for Saudi National Vision 2030
Ibtihal Ferwana · International Journal of Advanced Trends in Computer Science and Engineering · 2019
Twitter provides a valuable resource for opinion mining, which many applications can take place in investigating peoples' interests and concerns.In this paper, we are interested to investigate users' main concerns regarding the national Saudi vision of 2030.As the vision has many tracks of planning, we investigated the education and health sectors.The aim of this study is to show people's focus and interest in implementing this vision.For this, some unsupervised Machine Learning techniques are implemented.Three clustering algorithms are experimented (K-means, Latent Dirichlet Allocation (LDA) and Non-negative Matrix Factorization (NMF)).LDA and NMF give promising results, as both clustered the data into two significant discriminative clusters with a silhouette coefficient equals to 0.639 and 0.538 respectively.While K-means provides overlapped clusters with a silhouette coefficient equals to 0.181.Therefore, our models are able to cluster Arabic tweets in the context of the Saudi national vision.Additionally, they showed that the national vision 2030 is implemented in education more than in healthcare.