cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Browse apps to extend the software in the new JMP Marketplace
%3CLINGO-SUB%20id%3D%22lingo-sub-782064%22%20slang%3D%22en-US%22%20mode%3D%22CREATE%22%3EExiste-t-il%20un%20moyen%20d%E2%80%99indiquer%20%C3%A0%20LCA%20que%20les%20lignes%20sont%20class%C3%A9es%20par%20classe%26nbsp%3B%3F%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-782064%22%20slang%3D%22en-US%22%20mode%3D%22CREATE%22%3E%3CP%3EEnsemble%20de%20donn%C3%A9es%20typique%26nbsp%3B%3A%20120%20%C3%A0%20200%20lignes%2C%2015%20classes.%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CP%3EOn%20sait%20que%20les%20lignes%20sont%20class%C3%A9es%20par%20classe.%20On%20ne%20sait%20pas%20o%C3%B9%20se%20trouvent%20les%20%C2%AB%20barri%C3%A8res%20%C2%BB%20entre%20les%20classes.%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CP%3EEn%20l'%C3%A9tat%2C%20c'est-%C3%A0-dire%20sans%20tenir%20compte%20de%20l'ordre%20des%20lignes%2C%20LCA%20classe%20correctement%20environ%2094%20%25%20des%20lignes.%20Mon%20intuition%20est%20que%20si%20je%20savais%20comment%20dire%20%C3%A0%20l'algorithme%20que%20les%20lignes%20sont%20group%C3%A9es%20par%20classe%20en%20entr%C3%A9e%2C%20nous%20serions%20%C3%A0%20100%20%25.%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CP%3EVisuel%2C%20au%20cas%20o%C3%B9%20mon%20utilisation%20de%20%C2%AB%20group%C3%A9%20par%20%C2%BB%20ne%20serait%20pas%20claire%20%3A%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CP%3ELa%20bonne%20r%C3%A9ponse%20ressemblerait%20%C3%A0%20ceci%26nbsp%3B%3A%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CBLOCKQUOTE%3E%3CP%3E1a%3C%2FP%3E%3CP%3E1a%3C%2FP%3E%3CP%3E1a%3C%2FP%3E%3CP%3E1a%3C%2FP%3E%3CP%3E1a%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3C%2FBLOCKQUOTE%3E%3CP%3EToutes%20les%20lignes%201a%20seront%20toujours%20voisines%20en%20termes%20d'ordre%20de%20ligne%2C%20et%20toutes%20les%20lignes%201b%2C%20etc.%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CP%3EActuellement%2C%20j'obtiens%20parfois%20%3A%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CBLOCKQUOTE%3E%3CP%3E1a%3C%2FP%3E%3CP%3E1a%3C%2FP%3E%3CP%3E1a%3C%2FP%3E%3CP%3E%3CSTRONG%3E1c%3C%2FSTRONG%3E%3C%2FP%3E%3CP%3E%3CSTRONG%3E1c%3C%2FSTRONG%3E%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1b%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3CP%3E1c%3C%2FP%3E%3C%2FBLOCKQUOTE%3E%3CP%3EDes%20id%C3%A9es%20%3F%3C%2FP%3E%3CP%3E%20%3C%2FP%3E%3CP%3EPotentiellement%2C%20je%20pourrais%20cr%C3%A9er%20une%20colonne%20avec%20Row()%20comme%20valeur%2C%20mais%20je%20crains%20que%20la%20plupart%20des%20colonnes%20de%20clustering%20aient%20des%20valeurs%20de%20type%20%C2%AB%26nbsp%3Bclasse%26nbsp%3B%C2%BB.%20Pour%2015%20classes%2C%20chaque%20colonne%20utilis%C3%A9e%20peut%20avoir%20entre%202%20et%208%20valeurs%20uniques.%20Row()%20aura%20120%20valeurs%20uniques%2C%20ce%20qui%20donne%20l'impression%20d'ajouter%20un%20caract%C3%A8re%20g%C3%A9n%C3%A9rique%20au%20m%C3%A9lange.%3C%2FP%3E%3C%2FLINGO-BODY%3E%3CLINGO-LABS%20id%3D%22lingo-labs-782064%22%20slang%3D%22en-US%22%20mode%3D%22CREATE%22%3E%3CLINGO-LABEL%3EAnalyse%20et%20mod%C3%A9lisation%20de%20donn%C3%A9es%20de%20base%3C%2FLINGO-LABEL%3E%3CLINGO-LABEL%3EFusion%20et%20nettoyage%20des%20donn%C3%A9es%3C%2FLINGO-LABEL%3E%3C%2FLINGO-LABS%3E
Choose Language Hide Translation Bar
mtowle419
Level II

Is there a way to tell LCA that the rows are ordered by class?

Typical dataset: 120-200 rows, 15 classes.

 

Rows are known to be ordered by class. What is unknown is where the 'fences' between classes are.

 

As-is -- i.e., without taking row order into account -- LCA correctly classifies about 94% of rows. My intuition is that if I knew how to tell the algorithm that the rows are grouped by class on input, we'd be at 100%.

 

Visual, in case my use of 'grouped by' is unclear:

 

Correct would look like this:

 

1a

1a

1a

1a

1a

1b

1b

1b

1b

1b

1b

1b

1b

1c

1c

1c

1c

1c

1c

1c

All 1a rows will always be neighbors in terms of row order, and all 1b, etc.

 

 

Currently, I sometimes get:

 

 

1a

1a

1a

1c

1c

1b

1b

1b

1b

1b

1b

1b

1b

1c

1c

1c

1c

1c

1c

1c

Ideas?

 

Potentially, I could make a column with Row() as the value, but my worry is that most of the clustering cols have 'class'-type values. For 15 classes, each col in use might have between 2-8 unique values. Row() will have 120 unique values, which feels like throwing a wildcard into the mix. 

0 REPLIES 0