@David Mazziotti 

Kind help with how to get  the "table output" as usual given by BondDistances for the molecule got after adding the new bonds say.

I tried to add BondDistances(mol, bonding) to get bonddistance but got error

I get error 

Error, (in QuantumChemistry:-BondDistances) Set or list must have two items

One more is their anyway we can label the atoms in plot of the molecules for a verification purpose for any of my future coming molecules as by chance a bond between two atoms are missing and their are two or more atoms of same type to know the numbers to add the bond kind help.

Here i mean label with their indices 1 or 2 like that as they are unique 
 

@mmcdara 

Respected sir,

I tried to run the code with the Excel sheet as data the program gave error kind help to tweak code

Program with excel sheet command and errors it gave: TrainTest_and_ModelSelection_2.mw

Excel data sheet used:  analysis.xls

Kind help please to run with above data

I tried to somehow to import the excel data

Can we use some kind of database connectivity or create a dummy Excel sheet for the run and retrive as data becomes lager need some help to recode for that requirement

55 columns

Error, object too large in $

in Family_code

Please help please any help i will give thumbsup please help please

@tomleslie 

 table([(25, 50) = ["C-H", 1.08652243], (33, 34) = ["C-O", 1.36654528], (23, 48) = ["C-H", 1.08717708], (26, 30) = ["C-C", 1.40591714], (18, 23) = ["C-C", 1.39157249], (11, 16) = ["C-C", 1.39650421], (7, 36) = ["C-H", 1.08977154], (13, 18) = ["C-C", 1.39940202], (3, 6) = ["C-C", 1.47386499], (7, 11) = ["C-C", 1.47502068], (5, 6) = ["N-C", 1.39733282], (17, 23) = ["C-C", 1.39224064], (29, 31) = ["C-C", 1.40271915], (20, 46) = ["C-H", 1.08555101], (29, 52) = ["C-H", 1.08794255], (20, 25) = ["C-C", 1.38934841], (14, 20) = ["C-C", 1.40179385], (32, 55) = ["C-H", 1.08686798], (4, 8) = ["N-C", 1.39794063], (12, 17) = ["C-C", 1.39699284], (16, 42) = ["C-H", 1.08722951], (15, 41) = ["C-H", 1.08699816], (5, 9) = ["N-C", 1.45578467], (15, 21) = ["C-C", 1.41498728], (22, 47) = ["C-H", 1.08651829], (34, 35) = ["O-C", 1.42310435], (1, 2) = ["S-C", 1.78150330], (8, 13) = ["C-C", 1.40057453], (21, 27) = ["C-C", 1.40557782], (16, 22) = ["C-C", 1.39429839], (18, 44) = ["C-H", 1.08793888], (14, 19) = ["C-C", 1.40117130], (1, 3) = ["S-C", 1.72362728], (22, 27) = ["C-C", 1.39684144], (2, 5) = ["C-N", 1.37795936], (9, 14) = ["C-C", 1.52456453], (9, 37) = ["C-H", 1.09752950], (21, 26) = ["C-C", 1.48468482], (32, 33) = ["C-C", 1.39465444], (30, 53) = ["C-H", 1.08791452], (8, 12) = ["C-C", 1.39195618], (26, 29) = ["C-C", 1.40628233], (12, 39) = ["C-H", 1.08344312], (35, 58) = ["C-H", 1.09528489], (27, 51) = ["C-H", 1.08956367], (6, 10) = ["C-O", 1.22372628], (30, 32) = ["C-C", 1.40199679], (9, 38) = ["C-H", 1.08364570], (31, 54) = ["C-H", 1.08455521], (25, 28) = ["C-N", 1.34880132], (19, 45) = ["C-H", 1.08340943], (24, 28) = ["C-N", 1.34753442], (17, 43) = ["C-H", 1.08637148], (24, 49) = ["C-H", 1.08749437], (35, 57) = ["C-H", 1.09439070], (2, 4) = ["C-N", 1.29482084], (19, 24) = ["C-C", 1.39010000], (3, 7) = ["C-C", 1.34033764], (11, 15) = ["C-C", 1.40238725], (31, 33) = ["C-C", 1.39138564], (13, 40) = ["C-H", 1.08708279], (35, 56) = ["C-H", 1.09419651]])

All correct but

for this table when i excecute My graph drawn does not show the weights when i excecute can you help I dont know why

@Christian Wolinski How to store the output in a new list which list is it storing help with a sample code 

@mmcdara 

I have uploaded a excel sheet in the other thread can you help to run the code with that sheet 

For that thread I was getting error please help to verify the code with this the analysis excel sheet in the below thread

https://www.mapleprimes.com/questions/235524-Give-A-Data-For-Regression-In-Excel?reply=reply

@mmcdara 

Reply to the remider sir Please help the below

Please my kind help I have posted a reply to you two thread mainly 

Train test code thread and PLS PCR these as my excel sheet analysis  is not able to run in these codes please help

Kind help to Do PLS and PCR on the data say

analysis.xlsx

@mmcdara PLS+PCR_Linnerund.mw this i am not able to download dont know reason It says Page not found error

@Carl Love 

Kind help sir with Link to your KNN answer thread kind help

Please help 

@tomleslie 

The question was edited towards the end to generate a list of undirected graphs.

May be I need to use a for loop to convert all lists in the super list to a set of sets and add as graphs  to a new list.

This should work and short code help 

@mmcdara can you help with an Rcode for my steps

To run code train and test to get only  two or three regressor variables

If not get after a few runs it prints more data required

And if all satisfies the charts etc as I have 

Same algorithm 

@mmcdara I program should choose only a maximum of 2 or 3 independent variables only to design model if model R^2 is not more 90% it should print model is not good and it should print add more independent variables 

Here Y is the dependent variable and x1,x2,x3 etc are independent variables

In the actual data their could be more independent variables 

A_sample.xlsx

@mmcdara 

Step1; we need to import data from excel

Step 2: we split the data into train and test random

Step 3:

Here let us fix the number of regression variables or independent variables to be choosen at a time say K

I usual look for K to 2 or 3 in the model equation not more 

we need to apply stepwise regression or lasso regression that is we reduce the residual sum errors as told on the train data to get the best fit it gives 

Step 4: if the model best fits to predict the test data too then store the train and test data into Excel sheet

, And the predicted values of the dependent variable too based on the model equation obtained,

And give all the plots as suggested

With train data in one color and test In one Color with the fit model 

and stop

Or

If it doesn't we again run and randomly split the try to get the model

Hope train set 70% to test set 30%

Split

Or 

If no such model comes with two or three regressor variables even are certain number of number of runs

We print the need to create more regressor variables or independent variables to improve model

 A sample data 

In actual data their could be more independent variables

A_sample.xlsx

@mmcdara I want the fitting to pick the best of independent variables to get a good or best model for the dependent variable

It should not fit with all 50 variables and give a big regression equation

Will the above code be able to do that to

If not how can we pick the best subset of the independent variables which best fits the dependent variable from the data