Many algorithms you can used and compared for hierarchical clustering task on text data. I could recommend you to read the systematic literature review first, some of them are listed here:
- Hierarchical Document Clustering: A Review
- Document Clustering: A Detailed Review
And then you also used these algorithms:
- Katz-Classit Algorithm by Sahoo, et.al (Incremental Hierarchical Clustering of Text Documents in ACM CIKM 2016) (http://dl.acm.org/citation.cfm?id=1183667)
- Latest algoritms, Roy Model by Roy & Pokutta (Hierarchical Clustering via Spreading Metrics in NIPS 2016) (https://128.84.21.199/abs/1610.09269)
There are various text based clustering algorithms available....After applying clustering algorithms...you should have to measure the quality of your results using various indices which will give better quality of clusters.....But your quality of results also depends on the characteristics of your text datasets......You may check the below papers to get some initial ideas....
In my opinion, you can combine K-means and other hierarchical algorithms. K-means is better than other algorithms, but deciding k values, and the center of the k values are the hard problem. You can define k values by using hierarchical algorithms, then you can use k-means for your problem...
Sahoo, N., Callan, J., Krishnan, R., Duncan, G., & Padman, R. (2006, November). Incremental hierarchical clustering of text documents. In Proceedings of the 15th ACM international conference on Information and knowledge management (pp. 357-366). ACM.
I am afraid that it's not that simple: there is no algorithm that performs best: it will depend on your data set.
I suggest to use some software that can quickly run more than one algorithm, so you can test many and evaluate a comparison of results. For example, Weka is a simple software for this goal.
If this comment is useful, please appreciate it using the green arrow down here, thanks