"K-means clustering user specific cluster"

nachiket

Can I create clusters for each zone in k means clustering
i.e. One cluster for High one for low one for very high and one for Moderate

In normal k-means there is very much overlaping is there any way to minimize it

Sample Data:(1100 rows)
FID  Geology                                          Geomorphology                                              Land use_land cover    Rainfall          SLOPE    Soil                    zone
0  Fissile hornblende biotite gneiss  HIGHLY DISSECTED DIFLECTION SLOPE        FOREST                    1200-1400  >60%      BROWN CLAY          High
1  Fissile hornblende biotite gneiss  HIGHLY DISSECTED DIFLECTION SLOPE        FOREST                    1200-1400  30-60%  BROWN CLAY      Moderate
2  Charnockite                              HIGHLY DISSECTED DIFLECTION SLOPE    SEMI EVERGREEN    1200-1400    30-60%    BROWN CLAY    High
3  Charnockite                              HIGHLY DISSECTED DIFLECTION SLOPE  BUILDUP,RURAL        1200-1400    15-30%    BROWN CLAY    High
4  Charnockite                              HIGHLY DISSECTED DIFLECTION SLOPE  BUILDUP,RURAL        1200-1400    30-60%    BROWN CLAY    Very High
5  Charnockite                              LESS DESSECTED UNDULATING PLATEAU  AGRICULTURE        1600-2000    5-8%  ROCK OUT CROP    low