data_outliers3.m

我认为很不错的语音处理的ｍａｔｌａｂ源代码

% %
% %
% % ********************************************************************
% %
% %
% % data_outliers is written by Edward L. Zechmann
% %
% %  created 16 December    2007
% %
% % modified 20 December    2007    Added comments.
% %                                 Added col_name field.
% %                                 Improved ability to input data sets
% %                                 of cell arrays.
% %                                 Improved formatiting of output
% %                                 text file.
% %
% % modified  6  January    2007    Added features abs_rta, abs_other, sod
% %                                 and updated comments.
% %
% % modified 26 February    2007    Updated the row units and the array
% %                                 units.
% %                                 Replaced geomean with LMSloc robust
% %                                 mean.
% %                                 Modified code to print actual values
% %                                 including negative numbers, when
% %                                 the absolute value is used to process
% %                                 the descriptive statistics (mean, std,...).
% %
% % modified 19 September   2008    Updated Comments
% %
% % modified 16 December    2008    Added Arithmetic Mean.
% %
% % modified  9 January     2009    Added Default values.
% %                                 Added rounding the output.
% %                                 Added Analysis of Outliers.
% %
% % modified 10 January     2009    Finished updates.

% % modified 11 January     2009    Updated Comments
% %
% % ********************************************************************
% %
% % Please feel free to modify this code.
% %
% % See Also:  LMSloc, mean, std, t_confidence_interval
% %


% initialize abs_rta if empty
if nargin < 1 || isempty(abs_rta)
    abs_rta=0;
end

num_vars=length(varargin);

if nargin < 2 || isempty(dep_var) || ~isnumeric(dep_var)
    dep_var=1;
end

if ~isequal(dep_var, 1)
    dep_var=0;
end

if nargin < 3 || isempty(round_kind) || ~isnumeric(round_kind)
    round_kind=ones(1+num_vars, 1);
end

if length(round_kind) < (num_vars+1)
    for e1=(length(round_kind)+1):(num_vars+1);
        round_kind(e1)=0;
    end
end

if nargin < 4 || isempty(round_digits) || ~isnumeric(round_digits)
    round_digits=ones(1+num_vars, 1);
end

if length(round_digits) < (num_vars+1)
    for e1=(length(round_digits)+1):(num_vars+1);
        if isequal(round_kind(e1), 1)
            round_digits(e1)=3;
        else
            round_digits(e1)=0;
        end
    end
end


if nargin < 5 || isempty(abs_rta) || ~isnumeric(abs_rta)
    abs_rta=0;
end

if nargin < 6 || isempty(abs_other) || ~isnumeric(abs_other)
    abs_other=zeros(1+num_vars, 1);
end

% initialize abs_rta if shorter than varargin
if length(abs_other) < num_vars
    for e1=(length(abs_other)+1):num_vars;
        abs_other(e1)=0;
    end
end

if nargin < 7 || isempty(sod) || ~isnumeric(sod)
    sod=1;
end

% Value must be a scalar
sod=sod(1);
if  ~isequal(sod, 1);
    sod=0;
end

if nargin < 8 || isempty(num_std) || ~isnumeric(num_std)
    num_std=2;
end

if nargin < 9 || isempty(outfile) || (~ischar(outfile) && ~isnumeric(outfile))
    outfile='outliers_stats';
end

if nargin < 10 || isempty(save_file) || ~isnumeric(save_file)
    save_file=1;
end

if nargin < 11 || isempty(row_names)
    row_names='';
end

if nargin < 12 || isempty(row_unit)
    row_unit='';
end

if nargin < 13 || isempty(col_name)
    col_name='';
end

if nargin < 14 || isempty(array_names)
    array_names='';
end

if nargin < 15 || isempty(array_units)
    array_units='';
end


rta2=rta;
% For analyzing impulsive noise data
% num_data_rows would be the nnumber of frequency bands
% num_cols would be the number of microphone channels
[num_data_rows, buf]=size(rta);

% calculate the maximum number of columns
max_num_cols=1;
for e1=1:num_data_rows;
    if iscell(rta)
        max_num_cols=max(max_num_cols, length(rta{e1, :}));
    else
        max_num_cols=max(max_num_cols, length(rta(e1, :)));
    end
end


% initialize row_names if shorter than num_data_rows
if length(row_names) < num_data_rows
    if iscell(row_names)
        % If it is not a string assume it is an array of numbers
        if ischar(row_names{1})
            for e1=(length(row_names)+1):num_data_rows;
                row_names{e1}='';
            end
        else
            for e1=(length(row_names)+1):num_data_rows;
                row_names{e1}=0;
            end
        end

    else
        if ischar(row_names(1)) && ~isempty(row_names)
            buf=row_names;
            row_names=cell(num_data_rows);
            for e1=1:num_data_rows;
                row_names{e1}=[buf, ' ', num2str(e1)];
            end
        else
            for e1=(length(row_names)+1):num_data_rows;
                row_names(e1)=0;
            end
        end

    end

end


% initialize row_unit if shorter than num_data_rows
if length(row_unit) < num_data_rows
    if iscell(row_unit)
        % If it is not a string assume it is an array of numbers
        if ischar(row_unit{1})
            for e1=(length(row_unit)+1):num_data_rows;
                row_unit{e1}='';
            end
        else
            for e1=(length(row_unit)+1):num_data_rows;
                row_unit{e1}=0;
            end
        end

    else
        if ~isempty(row_unit) && ischar(row_unit(1)) && ~isempty(row_unit)
            buf=row_unit;
            row_unit=cell(num_data_rows);
            for e1=1:num_data_rows;
                row_unit{e1}=[buf];
            end
        else
            for e1=(length(row_unit)+1):num_data_rows;
                row_unit(e1)=0;
            end
        end

    end
end



% initialize col_name if shorter than max_num_cols
% if length(col_name) < max_num_cols
%     if iscell(col_name)
%         % If it is not a string assume it is an array of numbers
%         if ischar(col_name{1})
%            for e1=(length(col_name)+1):max_num_cols;
%                col_name{e1}='';
%            end
%        else
%            for e1=(length(col_name)+1):max_num_cols;
%                col_name{e1}=0;
%            end
%        end
%    else
%        if ischar(col_name(1)) && ~isempty(col_name)
%            buf=col_name;
%            col_name=cell(max_num_cols, 1);
%            for e1=1:max_num_cols;
%                col_name{e1}=[buf, ' ', num2str(e1)];
%            end
%        else
%            for e1=(length(col_name)+1):max_num_cols;
%                col_name(e1)=0;
%            end
%        end
%    end
%end

% col_name must be a string
if iscell(col_name)
    % If it is not a string assume it is an array of numbers
    if ischar(col_name{1})
        col_name=col_name{1};
    else
        col_name=num2str(col_name{1});
    end
else
    % If it is not a string assume it is an array of numbers
    if ischar(col_name)
        % Do nothing
    else
        col_name=num2str(col_name);
    end
end


% initialize array_names if shorter than num_vars=length(varargin)
if length(array_names) < (num_vars+1)
    if iscell(array_names)
        % If it is not a string assume it is an array of numbers
        if ischar(array_names{1})
            for e1=(length(array_names)+1):(num_vars+1);
                array_names{e1}='';
            end
        else
            for e1=(length(array_names)+1):(num_vars+1);
                array_names{e1}=0;
            end
        end

    else
        if ischar(array_names(1)) && ~isempty(array_names)
            buf=array_names;
            array_names=cell((num_vars+1), 1);
            for e1=1:(num_vars+1);
                array_names{e1}=[buf, ' ', num2str(e1)];
            end
        else
            for e1=(length(array_names)+1):(num_vars+1);
                array_names(e1)=0;
            end
        end

    end

end


% initialize array_units if shorter than num_vars=length(varargin)
if length(array_units) < (num_vars+1)
    if iscell(array_units)
        % If it is not a string assume it is an array of numbers
        if ischar(array_units{1})
            for e1=(length(array_units)+1):(num_vars+1);
                array_units{e1}='';
            end
        else
            for e1=(length(array_units)+1):(num_vars+1);
                array_units{e1}=0;
            end
        end

    else
        if ischar(array_units(1)) && ~isempty(array_units)
            buf=array_units;
            array_units=cell((num_vars+1), 1);
            for e1=1:(num_vars+1);
                array_units{e1}=[buf, ' ', num2str(e1)];
            end
        else
            for e1=(length(array_units)+1):(num_vars+1);
                array_units(e1)=0;
            end
        end

    end

end


ptsa=cell(num_data_rows,(num_vars+1));
nptsa=cell(num_data_rows,(num_vars+1));

mn_rt1a=zeros(num_data_rows,1);
mn_rt2a=zeros(num_data_rows,1);
stdrta=zeros(num_data_rows,1);
ci_inta=zeros(num_data_rows,1);
median_indexa=zeros(num_data_rows,1);
median_val=zeros(num_data_rows,1);
min_rta=zeros(num_data_rows,1);
max_rta=zeros(num_data_rows,1);

out_mn_rt1a=zeros(num_data_rows,1);
out_mn_rt2a=zeros(num_data_rows,1);
out_stdrta=zeros(num_data_rows,1);
out_ci_inta=zeros(num_data_rows,1);
out_median_indexa=zeros(num_data_rows,1);
out_median_val=zeros(num_data_rows,1);
out_min_rta=zeros(num_data_rows,1);
out_max_rta=zeros(num_data_rows,1);

% Initialize the maximum number of outliers in array rta data row.
max_num_out=0;

for e1=1:num_data_rows;

    if iscell(rta);
        buf1=rta{e1, :};
        if iscell(buf1);
            rt=buf1{1};
        else
            rt=buf1;
        end
    else
        rt=rta(e1, :);
    end

    num_cols=length(rt);

    if num_cols > 0
        if isequal(abs_rta, 1)
            rtb=abs(rt);
        else
            rtb=rt;
        end

        % Calculate the robust estimate of the mean.
        %
        % Statistical language for this type of mean is
        % "calculates the Least Median of Squares (LMS)