polyfits for continuous smile section with actual data

first question.

I got the data from the actual measurement, and I'm going to polyfit it with the data.

But I can do polyfits for the whole section, but I want to do polyfits multiple times in three minutes.

Second question.

The reason I used log (5080 - 430) is because I want to apply it as real data.

However, during P= polyfit (x, Y, 1), the b and m values were selected.

But I just want to do a regression analysis of the m values. How should I change the expression?

Can you help me?

% % 1번 위치
file_name = '20200207_1.xlsx'
x= linspace( 1/60, 349/60 , 349);
y = xlsread(file_name, 'sheet1','C6:C354' );
y_t=y';

figure(1);
plot(x,y_t,'ro','Linewidth',1,'Markersize',1);
grid; hold on; axis([0 7 0 6000]);

Y = log( y_t - 430);
P= polyfit(x, Y, 1);
b = log(5080 - 430)
m = P(1)

X1=linspace(1/60,349/60,2000);
Y1=(exp(m*X1+b))+430;
plot(X1,Y1,'b-', 'linewidth', 1);

xlabel ('t [h]')
ylabel ('indoor CO2 concentration [ppm]')
title('20200207 1location')
legend ('real time data','ach = 0.7273')
hold off;

[darova]

I run your code and got this:

Can you be more specific?

• But I can do polyfits for the whole section, but I want to do polyfits multiple times in three minutes.

Please show the entire section and how you want to use fit. Can you make a simple scheme or drawing?

[jungmin park]

First, I'll give you the basic knowledge of the equation I'm trying to do a regression.

is the concentration of co2 in Excel, C4:C354.

is the initial co2 concentration at the set time range (the caution is that the Cini value changes when the time range changes.)( is variable.) -- Second question.

is an outdoor co2 concentration, but we will use a fixed value of 430 ppm.

The unit of t is

If you organize the theoretical equation to do the first linear regression,

where

,

You need at least two actual data (2 points) to perform a regression based on the theoretical equation.

I used to do a regression analysis in the range from C4 to C354, but I try to do a regression analysis on two or three values. For example, C4 to C6 is a regression analysis with values, C7 to C9, and then continue to try a total of 117 times. -- first question.

I'll attach the picture to help you understand.

Additionally, it was successful if the x and y values were set to 100 random numbers between 0 and 1 and the primary linear regression was repeated 100 times. (I'll attach the code.)

But I don't know how to get it from Excel for the y and b (intercept) values that can change.

clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
clearvars;
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 24;
markerSize = 40;
numberOfTrials = 100;
for k = 1 : numberOfTrials
	x = randn(1, 100);
	y = randn(1, 100);
	% Fit a line.
	P = polyfit (x,y,1);
	m(k) = P(1);
	b(k) = P(2);
	fprintf('For trial #%3d of %3d, m = %.4f and b = %.4f.\n', ...
		k, numberOfTrials, m(k), b(k));
end
% DONE!  Now do the plotting.
% Plot m and b
hFig = figure;
subplot(2, 1, 1);
plot(m, 'b.-', 'LineWidth', 2, 'MarkerSize', markerSize);
grid on;
hold on;
plot(b, 'r.-', 'LineWidth', 2, 'MarkerSize', markerSize);
xlabel('Trial', 'FontSize', fontSize);
ylabel('m and b', 'FontSize', fontSize);
caption = sprintf('b and m for all %d points', numberOfTrials);
title(caption, 'FontSize', fontSize);
legend('m', 'b');
% Make a scatterplot of b vs. m but just for the indexes 95 to 100 (for some reason).
subplot(2, 1, 2);
scatter(m(95:100), b(95:100), markerSize, 'filled');
grid on;
xlabel('m', 'FontSize', fontSize);
ylabel('b', 'FontSize', fontSize);
title('b vs. m for the last 6 points (index 95 - 100)', 'FontSize', fontSize);
hFig.WindowState = 'maximized';

[darova]

Is this correct?

[jungmin park]

I think I'm a little wrong.

Because the function of P must be plotted against Y1 of theoretical equation with the m and b values defined by the expression below, but it seems to be plotted against Y value when looking at the picture.

-> I want to set the b value without estimating it as P(2).

I will attach the picture for your understanding.

I drew it rough. It's a curve plot for each section, but in the picture you've shown, it's a straight line.

[darova]

The question remains unclear for

You want to get coeffiecients for each region of a curve?

[jungmin park]

That's right. I want to get for each area.

For example, if the time interval is set to be 1 hour, the time zone is A4 : A63, A64 : A123, A124 : A183, A184 : A243, A244 : A303 ... The Cini value is set at 5122 ppm, C64 (2588 ppm), C124 (1436 ppm), C184 (909 ppm), and C244 (658 ppm).

In addition, I want to see the code of the graph that was first floating to me.