how to fit linear trend in time series

i have this time series liner fitted figure but i plotted this figure with the help of matlab tool box((basic fitting)(liner fit)) but i want to do same thing from matlab code.thank you in advance.

Answer 1

[Star Strider]

See if this does what you want:

% Create Data
yrs = linspace(1,100,100);
rainfall = 0.3*yrs + 350*rand(1,100);

% Fit Data
b = polyfit(yrs,rainfall, 1);
fr = polyval(b, yrs);

% Plot Data & Fit
figure(1)
plot(yrs, rainfall, '-b')
hold on
plot(yrs, fr, '-r')
hold off
h1 = legend('Data', 'Linear', 'Location','NE')
set(h1, 'FontSize',8)
[xlim ylim]
textposx = diff(xlim)*0.50+min(xlim);
textposy = diff(ylim)*0.95+min(ylim);
text(textposx,textposy, sprintf('y = %.2f*x %c %0.2f', b(1), char(45-(sign(b(1))+1)), abs(b(2))), 'FontSize',8)
xlabel('Years')
ylabel('Rainfall (mm)')

EDIT — Added textposx and textposy.

[devendra]

sir when i change something in code it mean when i inter my data in this code then equation does not show in the figure.sir my data size is rainfall_count(1x95).

% Create Data

%yrs = linspace(1,100,100);

yrs = linspace(1,95,95);

%rainfall = 0.3*yrs + 350*rand(1,100);

rainfall = rainfall_count;

% Fit Data

b = polyfit(yrs,rainfall, 1);

fr = polyval(b, yrs);

% Plot Data & Fit

figure(1)

plot(yrs, rainfall, '-b')

hold on

plot(yrs, fr, '-r')

hold off

h1 = legend('Data', 'Linear', 'Location','NE');

set(h1, 'FontSize',8)

text(50,350, sprintf('y = %.2f*x %c %0.2f', b(1), char(45-(sign(b(1))+1)),

 abs(b(2))), 'FontSize',8)

xlabel('Years')

ylabel('Rainfall (mm)')

[Star Strider]

I edited the code in my original Answer and added an adaptive position for the equation position. It should work without problems now.

[devendra]

Thank u so much sir this is working perfectly but little bit change in answer

y=0.04*x+10.80 wh

y=0.045*x+11 when i use matlab tool

why little bit difference in answer ?

[Star Strider]

My pleasure!

The MATLAB fitting function uses a different format in the presentation of the equation in the plot than I used.

The data are all available in full precision in my code. The slope is b(1) and the intercept is b(2).

[Muhammad Usman Saleem]

@Star Strider I have plot my timeseries plot through this method

myfile=load('time.txt');
xx = myfile(:,2);
dataa=myfile(:,1);
x=xx.';
ts1 = timeseries(x,1:length(x));
ts1.Name = 'Daily Count';
ts1.TimeInfo.Units = 'days';
ts1.TimeInfo.StartDate = '01-Jan-2003';     % Set start date.
ts1.TimeInfo.Format = 'mmm dd, yy';       % Set format for display on x-axis.
ts1.Time = ts1.Time - ts1.Time(1);        % Express time relative to the start date.
plot(ts1)

How can i use ploy fit against the time and ts1.Data? My data is a vector but date is in date format. Can you tell me how?

[Star Strider]

You need to convert the dates and times to date numbers using the datenum function. Use polyfit with 3 outputs so that it will do centring and scaling. Be sure to include all the outputs in your polyval call as well. Use the datetick function to plot the x-axis as the dates and times you want.

Answer 2

[dpb]

p=polyfit(y,x);

doc polyfit % for more details
doc polyval % to evaluate