how to create an index that keep tracking the hours in the data

Question

gjashta 2019-8-30

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/478302-how-to-create-an-index-that-keep-tracking-the-hours-in-the-data

评论： gjashta 2019-9-3

Is anyone helping me with a simple code ???

In the code below I want to find the peak load and average peak price in peak hour from 7-10 and then do it the same for each peak hours (11-15, 16-21 and 23-6am. I write the code below but I couldn't figure out how to calculate the peak load and average peak price for each peak with the same code

h=twomonth(:,1)
 p1= 7 <= h & h < 10
 price=twomonth(:,2);
 price=price.*p1  
 demand=twomonth(:,3);
 demand=demand.*p1
 Index1=(7:24:N)'; ss1=10-1
N1=length( Index1);
Data1=zeros(N1,2);
for k=1:N1
    kk=Index1(k); kkk=kk+ss1-1;
    dd=sum(d(kk:kkk));pp=sum(p(kk:kkk).*d(kk:kkk)); pp=pp/dd;
    Data1(k,:)=[pp dd];
end

2 个评论
显示无隐藏无

Star Strider 2019-8-30

编辑：Star Strider 2019-8-30

在 MATLAB Online 中打开

What does ‘isn’t working’ mean when you try to run it?

I can’t run it.

The error:

d=load.*p1
Error using load
Unable to read file 'matlab.mat'. No such file or directory.

Please do not name a variable ‘load’. It overshadows the load function, necessary for reading .mat files and other files.

What do you want to do in that line? Note that you are overwriting ‘d’ that you defined in the line just before it.

Also, ’p’, and ‘N’ are not defined.

gjashta 2019-8-30

Sorry Star Strider, you are right! I have fixed my mistakes in the code above.

Hope I was clear enough and you can help me!

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Guillaume 2019-9-2

2
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/478302-how-to-create-an-index-that-keep-tracking-the-hours-in-the-data#answer_390078

编辑：Guillaume 2019-9-2

在 MATLAB Online 中打开

Here's how I'd do it:

%twomonth: a Nx3 matrix, whose first column is hour, 2nd is price, 3rd is demand
bin = discretize(mod(twomonth(:, 1)-7, 24), [7, 11, 16, 22, 23, Inf]-7);  %offset bin edges by 7 in order to put the 23-6 in the same bin (5)
%bin == 1 is [7-11), bin == 2 is [11-16), bin == 3 is [16-22), bin == 4 is [22], bin == 5 is [23-6]
avgprice = splitapply(@mean, twomonth(:, 2), bin);
peakdemand = splitapply(@max, twomonth(:, 3), bin);
result = table({'7-10'; '11-15'; '16-21'; '22'; '23-6'}, avgprice, peakdemand, 'VariableNames', {'period', 'avgprice', 'peakdemand'})

I've included the lone 22 in the result, you can always remove row 4 if you don't want it.

You have to do a little bit of arithmetic as discretize can't give you the same bin number for the start and end of a range, but once you've got your bin numbers it's trivial to calculate your mean and peak with splitapply.

14 个评论
显示 12更早的评论隐藏 12更早的评论

Guillaume 2019-9-2

在 MATLAB Online 中打开

There's currently no day information in your matrix. If we assume a new day when the hours go from 23 to 0, then:

daybin = cumsum([1; diff(twomonth(:, 1)) < 0])

then, it's pretty much just the same, with an addition of a findgroups:

hourbin = discretize(mod(twomonth(:, 1)-7, 24), [7, 11, 16, 22, 23, Inf]-7);  %offset bin edges by 7 in order to put the 23-6 in the same bin (5)
hourranges = {'7-10'; '11-15'; '16-21'; '22'; '23-6'};
[bin, actualday, actualhour] = findgroups(daybin, hourbin);
actualhour = hourranges(actualhour);
avgprice = splitapply(@mean, twomonth(:, 2), bin);
peakdemand = splitapply(@max, twomonth(:, 3), bin);
weightedprice = splitapply(@(price, demand) mean(price .* demand), twomonth(:, 2), twomonth(:, 3), bin); 
result = table(actualday, actualhour, avgprice, peakdemand, weightedprice)

As for, "So hourRanges must be daily and another problem is that some days during the year has more/less hours due to daylight saving", that's vastly different from your original question and will require a different approach altogether. Currently, the necessary information is not in your data so I'm not sure what you want us to do about it. At the very least your data would need to contain date information (as datetime). A timetable would be more suitable for that instead of a matrix. We'd also need to know how you define the hour range for a particular day.

Guillaume 2019-9-3

在 MATLAB Online 中打开

Untested code, I don't have access to matlab right now:

%import the data. Use this or Adam's code
hourlydata = readtable('DataOrg.xlsx', 'ReadVariableNames', false);
hourlydata.Properties.VariableNames = {'date', 'hour', 'price', 'demand'};
%bin definition:
hourbins = [7, 11, 16, 22, Inf]; %bins must be contiguous and the last bin will wrap around to the start of the first bin
%discretize rows into bins:
[daybin, actualday] = discretize(hourlydata.date, 'day');
hourbin = discretize(mod(hour(hourlydata.hour) - hourbins(1), 24), hourbins - hourbins(1));  %the offset and mod is to wrap around the last bin
[group, groupday, grouphour] = findgroups(daybin, hourbin);
groupday = actualday(groupday);
starthour = hourbins(grouphour);
avgprice = splitapply(@mean, hourlydata.price, group);
peakdemand = splitapply(@max, hourlydata.demand, group);
weightedprice = splitapply(@(price, demand) mean(price .* demand), hourlydata.price, hourlydata.demand, group); 
result = table(groupday, starthour, avgprice, peakdemand, weightedprice)

Guillaume 2019-9-3

在 MATLAB Online 中打开

Oh yes, starthour probably needs transposing

result = table(groupday, starthour.', avgprice, peakdemand, weightedprice)

if it's not that then please give the size of each variables.

gjashta 2019-9-3

Thank you very much Guillaume! Now the code is OK!

请先登录，再进行评论。

how to create an index that keep tracking the hours in the data

2 个评论
显示无隐藏无

采纳的回答

14 个评论
显示 12更早的评论隐藏 12更早的评论

更多回答（0 个）

另请参阅

类别

标签

Community Treasure Hunt

how to create an index that keep tracking the hours in the data

2 个评论 显示 无隐藏 无

采纳的回答

14 个评论 显示 12更早的评论隐藏 12更早的评论

更多回答（0 个）

另请参阅

类别

标签

Community Treasure Hunt

2 个评论
显示无隐藏无

14 个评论
显示 12更早的评论隐藏 12更早的评论