Errata
This page contains the errata for the second edition. You can contact us via email if you want to report any errors.
Chapter 3
pg 74, Table 3.4: \(\mathbf{e}_{22}\) should be \(\mathbf{e}_{12}\)
Chapter 7
pg 208, Algorithm 7.2, line 7: in the denominator \(\sum_{i=1}^d \lambda_i\) should be \(\sum_{i=1}^n \lambda_i\)
Chapter 8
pg 230, fig 8.5: nodes labeled AC, CD and BCD should be shaded gray
pg 232, fig 8.6: nodes labeled AC, CD and BCD should be shaded gray
Chapter 10
pg 264, Algorithm 10.1, line 7: \(\mathbf{s} \in {\mathcal{C}}^{(k)}\) should be \(\mathbf{r} \in {\mathcal{C}}^{(k)}\)
pg 264, Algorithm 10.1, line 9: remove \(\mathbf{s}\) from \({\mathcal{C}}^{(k)}\) should be remove \(\mathbf{r}\) from \({\mathcal{C}}^{(k)}\)
pg 265, fig 10.2: nodes labeled C, AT, GT, TA, TG, AAA, AGA and AGG should be shaded gray. Also, extensions of an infrequent pattern are not shown.
pg 270, fig 10.4, part (f): label on the edge to leaf (1,5) should be \(T$\)
Chapter 12
pg 311, table 12.11: conv should be \(\infty\) for row 1, since conf is 1
pg 311, table 12.12: conv should be \(\infty\) for rows 1 and 3, since conf is 1
pg 317, Nonredundant Rules, Line 3: after \(sup(R)=sup(R')\) add provided that \(sup(W)=sup(X)\).
Chapter 13
pg 359, section on "Maximization Step", line 5: \(\ln E[P(\mathbf{D}\mathbf{\theta})]\) should be \(E[\ln P(\mathbf{D}\mathbf{\theta})]\)
pg 360, line 6 and line 15: \(\ln E[P(\mathbf{D}\mathbf{\theta})]\) should be \(E[\ln P(\mathbf{D}\mathbf{\theta})]\)
Chapter 15

pg 377, Algorithm 15.1, Line 17: if \(\mathbf{y} \in Core\) should be if \(\mathbf{y} \in Core\) and \(id(\mathbf{y}) = \emptyset\)
This will prevent an infinite loop.
Chapter 16
pg 395, last line: the weightedadjacency should be the weighted adjacency
pg 420, Line 10: A strongly connectedcomponent in a directed graph is defined a should be A strongly connected component in a directed graph is defined as a
Chapter 17
pg 434, contingency table: \(n=100\) should be \(n=150\)
pg 465, Q3, line 3: delete the entire line that reads: \(\mathcal{T}\) is completely determined by \(\mathcal{C}\)
Chapter 22
pg 549, eq 22.7: \(\frac{1}{k} \sum_{i=1}^r F_i\) should be \(\frac{1}{k} \sum_{i=1}^k F_i\)
pg 559, Random Classifier, Line 2: To see this think should be To see this, think
Chapter 23
pg 600, last line: matrix withi should be matrix with
Chapter 24
pg 630, line 5: 0.54 should be 0.56