The setting of my problem is as follows: we have a set of questions $T_1,..T_n$ and a set of workers $W_1,W_2,...,W_m$ and a matrix where a cell $c[i,j]$ is the answer of question $T_i$ by worker $W_j$. The domain of answers for each question is finite and discrete. I want to calculate the trustworthiness of the workers (whether I should trust his/her answers) and the correctness of the answers (the probability that an answer is correct). How can I model and solve this problem using EM?

According to wikipedia, there are three components of an EM algorithm: a set $X$ of observed data, a set of unobserved latent data or missing values $Z$ , and a vector of unknown parameters $\theta$, along with a likelihood function $L(\theta;X,Z)=p(X,Z|\theta)$.

In my opinion, the observed data $X$ is obviously the answers. But I don't know what is $Z$ $\theta$ and $L$.

Your Answer


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.