Please enable JavaScript.
Coggle requires JavaScript to display documents.
Survey "encuesta", Weighted ensemble (according accuracy),…
Survey "encuesta"
Dataset transformation and adaptation
153 rows; All binary columns (except **);
❌=Confounding variables
“subtopics”
(52)
Sport (4)
Sleep (6)
Food (4)
Vices (10)
Social (5)
Alien person (5)
Productive time (8)
Hobbies (10)
“pProfile”
(36 Columns)
❌Sex (3)
❌**Age, Size, Weight (3)
Actual status (3)
Home company (5)
About you (22)
“topics”
(8)
Sport
Sleep
Food
Vices
Social
Alien person
Productive time
Hobbies
K-means
Find best K
Elbow method
Silhouette method
Calinski-Harabasz index
Davies-Bouldin index
GridSearchCV
Fit K-means
Label the dataset
"topics_"
"subtopics_"
“pProfile_”
⠀
Analysis
Chi2_contingency
w\ Yates correction
3 more items...
ML Models
GridSearchCV
(Find best params
for each model)
5 more items...
“encuesta_”
Weighted ensemble
(according accuracy)
Reinsure importance
sum = 1
“{ }importance”
Ensamble with weight factor
“{ }importance”
Extract conclusions
Multiclass Multioutput
Classification
⠀
⠀
KFold (5 – Cross Validation)
and multiple iterations
(depending on
computational demand)
⠀
⠀
Extract characteristics importance
(importance sum = 1)
⠀
⠀