id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_3q7djukk75c2zkcpzjfjnkd7xu	Annie Louis	Which Step Do I Take First? Troubleshooting with Bayesian Models	2015	15	.pdf	application/pdf	8771	871	64	Specifically, we learn to assign complexity levels to solutions based on their linguistic makeup. original FAQ order using Kendall's τ (values are averages over 100 problem-solution sets). over the positions at which a solution with that complexity can occur and uses this distribution to order Qrm is the number of times a solution with position r is given complexity m over the full corpus. Table 3: Most likely words assigned to varying complexity levels by the permutations-based model. Specifically, we can compute the expected complexity of the solution set for problem i, using the inferred distribution over levels θi: In the previous section, we showed how our models can assign an expected complexity value to a solution text or an entire problem. the ranking model on 240 problem-solution sets; level 1 will be ordered after the solution with complexity 2.	./cache/work_3q7djukk75c2zkcpzjfjnkd7xu.pdf	./txt/work_3q7djukk75c2zkcpzjfjnkd7xu.txt