Implement all nearest neighbor

Lets say I want to find all neighbors. A function that will take in an excel file with data of different cars (caryear carmileage carhorsepower carlife) amd I want to be able to input data of a car I want and if it already exist according to my input then return that but if it does not return all the colosest/similar version of that car according to some all nearest neighbor implemantation. Not sure if that makes sense but I would appreciate some help :)

[Mathieu NOE]

hello again

is this a kind of extension of your previous post ?

User input compared with excel sheet data and returns best fit - (mathworks.com)

[ML]

hello yes.

[Image Analyst]

OK, but did @bransa's Answer below work?

[ML]

No I dont think I explained it well. Basically given a list of variable names find nearest neighbor.... consider min/max ranges. How can I do all this using knnsearch?

Answer 1

[bransa]

I would use logical indexing to query your excel data. When your return is empty, then prioritize the dimensions you want to search by and either use a combination of min-abs(difference) (one item searched) or knnsearch for a list of queries. Here is something to show what I am getting at. Not sure it runs bc I don't have sample data

found_index = find(carYear == inputYear & ...
    carMileage == inputMileage & ...
    carHorsepower == inputHorsepower & ...
    carLife == inputLife);

if ~isempty(found_index)
    fprintf('%s meets your input specifications.\n',carName)
else
    % first try to find the year & horsepower ignoring the mileage and life
    % just because this makes sense to me
    found_index = find(carYear == inputYear & carHorsepower == inputHorsepower);
    if ~isempty(found_index)
        if length(found_index) == 1
            fprintf('%s is the same year & horsepower.\n',carName)
        else
            fprintf('the following cars are the same year & horsepower:\n')
            % now you have more than one car that meets 2 conditions. 
            % find the min difference for mileage and life 
            
            [~,minLifeIndex] = min(abs(carLife(found_index)-inputLife(found_index)));
            [~,minMileageIndex] = min(abs(carMileage(found_index)-inputMileage(found_index)));
            % note: your min~Index refers to the index of the
            % array found_index. The index that meets the criteria in
            % carMileage is carMileage(found_index(min~Index))
            
            fprintf('\tclosest in life: %s\n',carName(found_index(minLifeIndex)))
            fprintf('\tclosest in mileage: %s\n',carName(found_index(minMileageIndex)))
            
        end
        
    else
        % some other less rigorous set of conditions
    end
end

[ML]

How could I implement knnsearch here?