46528

Instmctions on Using the tool

(Building a prediction Model)

Step I Enter Your D sta

(A Enter yo u r data. in The O ata w or ks he et. s tartin g fro m th e ce II AC 3D5

(B i The obs ervations s hould be in rows and the yariables s hould be in columns.

(C Above each column, choose appropriate Type(Omit. Output. Cont. Cat)

To drop a column from model - s et the type = Omit Totreatacolumn a; categoricaJ Input s ettype = Cat To treat a column as continuous Input s ettype = Cont Totreatacolumn as Output. settype = Output

You can hswe atmostlD output yariables. Application will automaticaJlytreatthem all as contini Usually one builds prediction model with 1 output only.

If you have s ay.2 outputvariables Y1 and Y2. both of which depend on the s ame s et of Input v< you may be better off. building 2 s eparate modeb - One with Y1 as Output. another one with Y2

You can hat/e at most5Q input yariables. out of which atmost40 could be : otegoricaL

Make suręthatthe number oflnput(Cat& Cont) columns exactly match with the number enter>

(D PIease make surę thatyour datadoes nothave blank rows or blank columrc.

( E) Cont inu ous Inputs:

Any noty nu mb er in Cont column w ill be treated as miss ing vaJue.

Application will replace it by the column mean

(E) Categorcallnputs:

Any blank celi orcelb containing Excel error in Cat column will be treated as missinę Application will reaplce it by the most frequently occuring category.

Categoiy labeb are • in; otl - lables good. Good. GoOd. GOOC will all betn There s hould be at ko.- t2 ot; ory anons in each category of a Cat column.

If one of the categoiy of a C at column has oni/ 1 obs eivation. you s hould do one of Remove that observation OR

Rename the category to any other categories ofthat Cat column.

Step 2: Fiłł up Model Inputs ( Ai Fili up the model inputs in the Us er Input Page.

(B Make surę thatyour inputs a/e with in the rangę ofyaJu es allowed by the application.

(C ; Click the 'Build Modef button to start model ing.

Step 3: Resułtsof fdodełing

(A ANeuraJ Network model is basical^ asetof weights betweenthe Isyers ofthe net.

At the end ofthe run. the finał s et of weights aresaved inthe Calcsheet (Bi The output page ofthis file will show you thevalues of MSE and ARE on thetraining and va as the training ofthe model progress es. Two charts s howing training and Validation MSE’s ha\«* be en already provided inthe Outputsheet

(C i In Us erlnput page ifyou h»/e as ked to s ave the model in as eparate file. then a new file w ill be created containing the model inputs. your data, and the fitted model ( i.e. the weights) You will be able to us e thts file as a calculator to do prediction. giyen any new input.

Step 4: StudyProfiles

Fitted model is asurface inp-dimension wherethe number of your inputs is p.

Unless p ts 2 or less. it is not possibletoshowthesurface graphicaJly.

Profile plot ts the next best wsy tovisuali2ethts fitteds urface.