Showing posts with label chebyshev inequality. Show all posts
Showing posts with label chebyshev inequality. Show all posts

Sunday, July 26, 2009

Chebyshev's Inequality to Roughly Estimate Area Under an Unknown Probability Density Curve: Reposted from www.SixSigmaPractice.multiply.com

I have always been fascinated by this generalized area under the curve theorem.
Though the estimate in itself is very weak, Chebyshev's inequality provides a very strong statistical basis for control charting.
Here is a utility that compares the actual area under the curve against to that estimated by Chebyshev's inequality.


The JMP JSL Script




Clear Globals();
new=EXPR(dt=open());
use_curr=expr(dt=Current Data Table());
if(is empty(Current Data Table()),new,use_curr);
bound=1.5;
j = 1;
col_list = {};
For( j = 1, j <= N Col( dt ), j++,
col_list[j] = Column( j ) << get name
);


COMM_RUN = Expr(
col = Column( col_name );
modelling=col<<get modeling type;
CONT=EXPR(
val1 = col << get as matrix;
counter = N Rows( val1 );
T_counter = 0;
mu = Mean( val1 );
sigma = Std Dev( val1 );
i = 1;
For( i = 1, i <= counter, i++,
If( val1[i, 1] <= mu + bound * sigma & val1[i, 1] >= mu - bound * sigma,
T_counter = T_counter + 1;
i = i + 1;
,
i = i + 1
)
);

Actual = T_counter / counter * 100;
T_estimate = (1 - (1 / (bound ^ 2))) * 100;
);
NON_CONT=EXPR(Throw ("Data Column does not contain a continous variable data"));
if(modelling=="Continuous",CONT,NON_CONT);
);

COMM_PRINT = Expr(
PRINT_TEXT = "Actual area inside mean +/- " ||char(bound) || "*sigma is " || Char( round(Actual,2) ) || " while Chebyshev's inequality estimate is " || Char( round(T_estimate,2) );
PRINT_VALUES = "Mean is equal to " ||char(round(mu,3)) || ". Sigma is equal to " ||char(round(sigma,3)) || ". The bounded interval is equal to " ||char(round(mu-bound*sigma,3)) || " to " ||char(round(mu+bound*sigma,3)) || ".";
Print( PRINT_VALUES );
Print( PRINT_TEXT );
);


COMM_OUTPUT = Expr(
OUTPUT_REPORT = New Window( "Results",
Border Box(
Left( 10 ),
Panel Box( "Simulation Result for " || char(col),
Text Box(" "),
V LIST BOX(
Text Box( "" ),
Text Box( PRINT_VALUES ),
Text Box( " " ),
Text Box( PRINT_TEXT ),
Text Box(" "),
Text Box( "sixsigmapractice.multiply.com" )
)
)
)
)
);

New Window( "Interface",
Border Box(
Left( 10 ),
Panel Box( "Set values and click on RUN",
V List Box(
V List Box(
H List Box(
Text Box( "Set Bound = k" ),
Text Box( " " ),
combo_list = Combo Box(
{"1.5", "2", "2.5", "3", "3.5", "4", "4.5", "5", "5.5", "6"},
COMBO_COMM = Expr(
k = combo_list << getselected;
bound = Num( k );
);
COMBO_COMM;
),

),
Text Box( " " ),
H List Box(
Text Box( "Select Column" ),
Text Box( " " ),
col_select = Combo Box( col_list ),
COL_COMM = Expr( col_name = col_select << get );
COL_COMM;
)
),
Text Box( " " ),
Panel Box( "Command Button",
Button Box( "RUN ESTIMATE",
COMBO_COMM;
COL_COMM;
COMM_RUN;
COMM_PRINT;
COMM_OUTPUT;
OUTPUT_REPORT << reshow;
OUTPUT_REPORT<<move window(250,50)
)
)
)
)
)
);

Saturday, July 11, 2009

Extending Your JMP's Capability Thru Scripting

If your are a JMP user do you know that you can increase your JMP's capability by leaps and bounds through scripting?

Yes you can! I do create scipts which utilities range from an ordinary alarm clock, mail sender, matrix solver, to report generator, and so on. Some of my work are shared online and can be found here:
http://sixsigmapractice.multiply.com/

http://www.sas.com/apps/demosdownloads/jmpFileExchange_PROD__sysdep.jsp?packageID=000416&jmpflag=Y&searchvar=userName&searchval=Reynald%20Francisco

In my multiply account, i do post many JSL scripts. In here i will try to foucs more on teaching how to script.

Till next entry!
 
Custom Search