BLIM and Bayesian Assessment¶
The Basic Local Independence Model¶
The BLIM (Doignon & Falmagne, 1999, Ch. 12) is a probabilistic model for response patterns on a knowledge structure.
Parameters¶
For each item \(q\):
\(\beta_q\) (slip): \(P(\text{incorrect} \mid q \in K)\) — the probability of a careless error on a mastered item.
\(\eta_q\) (guess): \(P(\text{correct} \mid q \notin K)\) — the probability of a lucky guess on an unmastered item.
The constraint \(\beta_q + \eta_q < 1\) ensures identifiability: a mastered item is more likely to receive a correct response.
Likelihood¶
The probability of a response pattern \(R\) given a knowledge state \(K\):
where:
The local independence assumption means responses to different items are conditionally independent given the state.
Bayesian assessment¶
Prior¶
Initially, each state is equally likely (uniform prior):
Update¶
After observing a response \(R_q\) on item \(q\):
This is a standard Bayes update, applied sequentially for each observed response.
Item selection: Expected Information Gain¶
The EIG criterion selects the item that maximally reduces entropy:
where \(H\) is Shannon entropy and the expectation is over both possible responses, weighted by their marginal probability.
Items vs instances¶
In practice (e.g., ALEKS), each item is a problem type and has multiple instances — concrete questions of equivalent difficulty. The assessment engine:
Computes EIG at the item level.
Selects a random un-asked instance of the best item.
Updates the posterior on the parent item.
This distinction is theoretically grounded: instances are treated as equivalent under the model — they test the same latent competency with equal difficulty (Cosyn et al., 2021).
Parameter estimation (EM)¶
When \(\beta_q\), \(\eta_q\) are unknown, they can be estimated from data via the EM algorithm:
E-step: For each response pattern \(R_r\) and state \(K_k\):
M-step: Re-estimate parameters from weighted sufficient statistics:
The log-likelihood is guaranteed to increase at each iteration.
References¶
Doignon, J.-P., & Falmagne, J.-C. (1999). Knowledge Spaces, Ch. 12.
Cosyn, E., et al. (2021). A practical perspective on knowledge space theory: ALEKS and its data. Journal of Mathematical Psychology, 101.
Heller, J., & Wickelmaier, F. (2013). Minimum discrepancy estimation in probabilistic knowledge structures.