1: \section{\label{sec:ana}Analysis}

2: During 1,489 days of livetime corresponding to 92 kton$\cdot$year of exposure,

3: 12,179 events are recorded after the first stage reduction process.

4: Of these, 8,207 are identified as single ring.

5: 

6: \par

7: It has been well known for some time that there is a deficit of

8: events induced by atmospheric muon neutrino oscillations \cite{skosc}.

9: In order to estimate accurately the actual number of background events,

10: the atmospheric neutrino flux is normalized based on the observed deficit of

11: $\mu$-like events due to well established neutrino oscillation phenomena and 

12: the observed number of $e$-like events.

13: The normalizing factor is determined as multiplying 0.67 for charged current 

14: $\nu_{\mu}$ events and 1.07 for charged current $\nu_e$ and all neutral

15: current events.

16: Finally we normalize the 100 year MC sample (2,246 kton$\cdot$years) to the 

17: 92 kton$\cdot$year sample of \sk exposure.

18: Based on these simulated signal and background events,

19: we determine selection criteria for each nucleon decay mode,

20: which is discussed in the following sections.

21: All detection efficiencies given below include all relevant branching ratios.

22: 

23: \input{nukp}

24: \input{nukz}

25: \input{mukz}

26: \input{ekz}

27: \input{systematicerror}

28: \input{results}

29: 

30: