Comparative analysis of supervised and unsupervised learning algorithms for online user content suicidal ideation detection