Location | Correction |
---|---|

Chapter 4, pg. 110, (PageRank, Figure 4.11) | Line 14 should read "the set of pages such that" Line 21 should read "Rq <- Rq ..." instead of "Rp <- Rq ..." |

Chapter 5 (Pseudocode) | Many of the pseudocode figures in Chapter 5 have been clarified. See the slides for Chapter 5 for updated versions. |

Chapter 5 (Elias-Gamma codes), page 147 | k_dr should be defined as (k_d + 1) - 2^{\floor{\log_2(k_d+1)}} |

Chapter 5 (Elias-Gamma codes), page 149 | k_d should be defined as ceil(log k - 1). If k > 1, then the leading bit will be 1. |

Chapter 5, page 150 | 1,2,1,6,1,2,6,11,... should be 1,2,1,6,1,3,6,11 |

Chapter 5, page 150 | 81,82,81,86,81,82,86,... should be 81,82,81,86,81,83,86,... |

Chapter 5, Figure 5.17, page 168 | In the bottom row the third final score should be 3:1 rather than 2:2. |

Chapter 6, pg. 199 | "Other studies of have" should be "Other studies have" |

Chapter 7, pg. 248, (Table 7.1, Total column) | (N - r_i) should be (N - n_i) |

Chapter 8, pg. 317 (Table 8.4) | The 40% recall column should have an average of 0.55 (rather than 0.47). |

Chapter 8, pg. 317 (Figure 8.6) | The 40% recall point should be at 0.55, but appears to be at 0.60. |

Chapter 9, pg. 376 (Divisive clustering, Figure 9.11) | Revised figure |

Chapter 9, pg. 382 (K-Means) | "As the algorithm proceeds, the number of clusters increases." - should be "decreases" |

Chapter 10, pg. 435 (Rating with user clusters) | The u variable in the correlation formula should have an overline (mean), instead of a hat. |