1 \documentclass[a4paper]{article
}
8 \title{Exercise: Machine Learning
}
14 \subsection*
{Chapter
5: Machine Learning
}
19 Method & Correctly classified & Root relative squared error\\
21 NaiveBayes & $
96.6449\%$ & $
35.7222\%$\\
22 NaiveBayes (
10FCF) & $
96.4352\%$ & $
37.1926\%$\\
23 J48 & $
96.6449\%$ & $
34.9136\%$\\
24 J48 (
10FCF) & $
96.4352\%$ & $
36.5122\%$\\
27 \caption{Results for
\texttt{P1D
} and
\texttt{FD
}}
32 \begin{tabular
}{lllllll
}
34 \texttt{P2D
} &
\texttt{P1D
} &
\texttt{N2D
} &
\texttt{N1D
} &
\texttt{FW
} & Correctly classified\\
36 \checkmark{} &
\checkmark{} &
\checkmark{} &
\checkmark{} &
\checkmark{} & $
98.3225\%$\\
37 \checkmark{} &
\checkmark{} &
\checkmark{} &
\checkmark{} & & $
97.3905\%$\\
38 \checkmark{} &
\checkmark{} & &
\checkmark{} &
\checkmark{} & $
98.5555\%$\\
39 \checkmark{} &
\checkmark{} & &
\checkmark{} & & $
97.507\%$\\
40 \checkmark{} &
\checkmark{} & & &
\checkmark{} & $
98.0429\%$\\
41 \checkmark{} &
\checkmark{} & & & & $
95.5732\%$\\
43 &
\checkmark{} &
\checkmark{} &
\checkmark{} &
\checkmark{} & $\%$\\
44 &
\checkmark{} &
\checkmark{} &
\checkmark{} & & $\%$\\
45 &
\checkmark{} & &
\checkmark{} &
\checkmark{} & $\%$\\
46 &
\checkmark{} & &
\checkmark{} & & $\%$\\
47 &
\checkmark{} & & &
\checkmark{} & $
98.2992\%$\\
48 &
\checkmark{} & & & & $
96.6449\%$\\
50 & &
\checkmark{} &
\checkmark{} &
\checkmark{} & $\%$\\
51 & &
\checkmark{} &
\checkmark{} & & $\%$\\
52 & & &
\checkmark{} &
\checkmark{} & $\%$\\
53 & & &
\checkmark{} & & $\%$\\
54 & & & &
\checkmark{} & $
88.4436\%$\\
61 \subsection*
{Chapter
6: Exercises
}
63 \item\emph{If we look at the Viterbi algorithm, we see that the
64 probability of state at a given position is calculated $n$ the basis of
65 the preceding $n$ states. However, it is claimed that the algorithm
66 takes into account the whole sequence. Explain in your own words (at
67 most $
100$) how the probability is influenced by the rest of the
68 sequence, i.e.\ both the positions more than n back and the following
71 \item\emph{Explain in your own words (at most
50) how the EM algorithm
72 works. I don't mean the mathematics, but the underlying concept.
}