Read Me First

1. To run the experiments on the Perplexity performance.

	test_perplexity(paras,control,methods);
	Input: paras and control: load from CoraParas.mat or CiteSeerParas.mat
	       methods: A string array specifying the methods: "LINK_LDA" is for Link-LDA, "LCA" is for LDA model, and "LCA_Base" is for BPT model.
	       
	Output: the result is written in a mat file.
	
	
2. To run the experiments on the Perplexity with different \beta value in the BPT model.

	ParaSelectionPerplexity(paras,control,methods);
	

3. To detect the latent topics from the corpus
	
	test_topics(paras,control,methods);
	Input:  
		methods: "LINK_LDA" is for Link-LDA model, "test_LCA" is for LDA model, and "LCA_Base" is for BPT model
	Output: the result is written in a matfile and txt file
	
4. To detect the topic trends
	
	topic_trend(paras,control,methods);
	
	Output: the result is written in a txt file.
	
	

