Weather Data: Complete and Ward Linkages
The last chart was good, but if you remember the K-Means and K-Medoids algorithms results, you may remember that there was at least one more line that unlike all the others goes downwards close to July. The average linkage in hierarchical clustering didn't catch that dynamic.
We saw that for complete and ward linkages there is sense to consider 4 clusters. Let's find out will they catch that?
Swipe to start coding
- Import
numpywithnpalias. - Iterate over the
linkageslist. At each step:
- Create a hierarchical clustering model with 4 clusters and method
jnamedmodel. - Fit the numerical data of
tempand predict the labels. Add predicted labels as the'prediction'column totemp. - Create a
temp_resDataFrame with monthly averages for each group. To do it group the values oftempby the'prediction'column, calculate themean, and then apply the.stack()method. - Add column
'method'totemp_resDataFrame with valuejbeing repeated the number of rows intemp_restimes. - Merge
resandtemp_resdataframes using.concatfunction ofpd.
- Reassign the column names of
resto['Group', 'Month', 'Temp', "Method"]. - Within the
FacetGridfunction set thecolparameter to'Method'. This will build a separate chart for each value of the'Method'column. - Within the
.mapfunction set theseabornline plot function as the first parameter.
Løsning
Tak for dine kommentarer!
single
Spørg AI
Spørg AI
Spørg om hvad som helst eller prøv et af de foreslåede spørgsmål for at starte vores chat
Opsummér dette kapitel
Explain code
Explain why doesn't solve task
Fantastisk!
Completion rate forbedret til 3.57Weather Data: Complete and Ward Linkages
Stryg for at vise menuen
The last chart was good, but if you remember the K-Means and K-Medoids algorithms results, you may remember that there was at least one more line that unlike all the others goes downwards close to July. The average linkage in hierarchical clustering didn't catch that dynamic.
We saw that for complete and ward linkages there is sense to consider 4 clusters. Let's find out will they catch that?
Swipe to start coding
- Import
numpywithnpalias. - Iterate over the
linkageslist. At each step:
- Create a hierarchical clustering model with 4 clusters and method
jnamedmodel. - Fit the numerical data of
tempand predict the labels. Add predicted labels as the'prediction'column totemp. - Create a
temp_resDataFrame with monthly averages for each group. To do it group the values oftempby the'prediction'column, calculate themean, and then apply the.stack()method. - Add column
'method'totemp_resDataFrame with valuejbeing repeated the number of rows intemp_restimes. - Merge
resandtemp_resdataframes using.concatfunction ofpd.
- Reassign the column names of
resto['Group', 'Month', 'Temp', "Method"]. - Within the
FacetGridfunction set thecolparameter to'Method'. This will build a separate chart for each value of the'Method'column. - Within the
.mapfunction set theseabornline plot function as the first parameter.
Løsning
Skift til skrivebord for at øve i den virkelige verdenFortsæt der, hvor du er, med en af nedenstående mulighederTak for dine kommentarer!
single
