interactively extract, visualize, and rank features from measured or simulated data for machine diagnostics and prognostics -凯发k8网页登录

initiate a new app session by importing source data into the app from your matlab workspace. you can import data from tables, timetables, cell arrays, or matrices. you can import data from a single source that combines the data of multiple ensemble members or import individual ensemble members from separate sources. you can import labeled signal sets created in signal processing toolbox™. you can also import an ensemble datastore that contains information that allows the app to interact with external data files. your files can contain actual or simulated time-domain measurement data, spectral models or tables, variable names, condition and operational variables, and features you generated previously. diagnostic feature designer combines all your member data into a single ensemble data set. in this data set, each variable is a collective signal or model that contains all the individual member values.

for more information about importing data, see .

for more information about terms related to data ensembles, see more about.

for more information about organizing your data for import into the app, see .

generate and rank a feature set automatically using auto features. when you select one or more signals or spectra, auto features computes a predefined set of features that are appropriate for the variable type. the automatic computations include:

for more information, see .

specify default plotting options for all plots that you generate during your app session. you can set these options before you generate your first plot, or at any time during your session. new settings apply only to plots that you generate after setting the options and not to plots that you generated earlier. you can temporarily override the plot options settings for individual plots without changing your specified defaults for subsequent plots. when you click plot options, you open a dialog that allows you to set options in the following panes.

generate a plot of an ensemble variable or feature table. to generate a plot, first select a variable or feature table from the data browser. the plot gallery shows icons for the compatible plot types. the following table describes the plot types for each type of selection.

specify a frame policy when you want to perform data processing on sequential segments of a signal rather than the full signal at once. a frame policy consists of a frame size and a frame rate. the frame size is the interval over which the frame data is collected. the frame rate is the time interval between frame start times.

for more information on frame-based processing, see .

specify options when you want to modify one or both the following settings.

select options for processing your data into new signals. use these new signals as inputs to other processing options or as inputs to feature generation. most processing options operate on each ensemble member. you can also perform ensemble-level processing to view how the ensemble behaves as a whole. each option selection opens a new tab for your specifications. selection of an option also opens a general data processing tab if that tab is not already open. the data processing tab provides information about the input signal.

to specify a signal to process, select a variable from the variables pane prior to selecting the data processing option. to change the signal after opening the option tab, close the option tab and select a new signal either in the variables pane or from the signal list in the data processing tab.

for more information about the processing options and the parameters that you can set for each option, see:

compute features in the time domain. signal features apply to any signals. time series features are features extracted from stationary time series. model-based features are features extracted using autoregressive (ar) models. rotating machinery features are specialized metrics that apply to gearing. nonlinear features provide metrics that characterize chaotic behavior in vibration signals. custom features are features that you define by adding custom functions to the app. you can add existing matlab functions or create new functions using a template and then generate and rank the features alongside built-in app features

to specify a signal source for your features, select a signal variable from the variables pane prior to selecting the time-domain features option. to change the signal after opening the option tab, close the option tab and select a new signal either in the variables pane or from the signal menu in the time-domain features tab.

for more information about time-domain features options and the parameters that you can set for each option, see:

compute features in the frequency domain. spectral features are general metrics that apply to any spectrum, such as the peak amplitude across the full specified frequency range. bearing faults features, gear mesh faults features, and custom faults features are specialized metrics for rotating machinery that focus on spectral behavior within specific fault bands that bound characteristic frequencies, where faults can occur, of the components of the system. custom features are features that you define by adding custom functions to the app. you can add existing matlab functions or create new functions using a template and then generate and rank the features alongside built-in app features.

for more information about the frequency-domain features, see

open the feature ranking tab to perform supervised, unsupervised, or prognostic ranking for the feature table that you select. for more information, see feature ranking tab.

export features, or your entire data set, to use them or share them outside of the app. generate code to reproduce your feature computations in a matlab function or a simulink block.

for more information about the export options, see:

use the panner to focus on data segments in the x-axis range that you specify and to change the plot scale. the panner provides a strip plot beneath the main plot. to focus on a section of the main plot, move the handles. to change the scale of the plot, select one of the options in scale.

use the options in the first column of the view section to override the defaults in the plot options specification. the available options vary by plot type. when you change these settings in the plot tab, you change them only for the current plot. for more information on these options, see plot options.

use normalize y axis when you are plotting multiple variables and want to view the variables on the same [-1, 1] scale. the relative signal amplitudes within a variable do not change.

in a signal or spectrum plot, you highlight an individual member by positioning your cursor on the member trace. select show signal information to display both the variable member that you highlight and the condition label for that member in the lower right corner.

if you select data cursors, show signal information also displays the distance between the two cursors. for more information, see data cursors.

specify how to plot multiple variables together.

select data cursors to display values of key points in your signal. data cursors are horizontal and vertical bars that you position over a point of interest, such as a peak value. the cursors display the x and y positions. to display the distance between the cursors, select show signal information. to lock the bars so that they move together, select one of the lock spacing options.

click select features to open a selectable list of features to plot. use select features, for example, when you have generated many features but you want to focus on a subset in a single plot panel.

for more information about selecting features, see .

select the condition variable to base feature histograms on. the feature histograms use color to visualize the separation of data groups with different labels for that condition variable.

example: faultcode

specify histogram resolution using bin width, bin method, number of bins, and bin limits. the bin settings apply to all the histograms for the feature table.

the bin settings are not independent. the app histogram algorithm uses an order of precedence to determine what to use:

select a supervised classification ranking method to assess how effectively each feature separates data with different condition labels. if you have already ranked your features, you can rank again with a different method and display the resulting rankings together. each method uses a different statistical approach.

the menu differentiates between two-class and multiclass ranking methods.

the default ranking method for two-class condition variables, t-test, is the simplest method, as it only considers whether the means of the two labeled groups are equal or not. t-test is primarily useful for identifying ineffective features to discard.

the table lays out the influence of specific criteria on ranking-method selection.

selecting a method activates a new tab with a name that matches the ranking method. for more information on this method-activated tab, see ranking method tabs.

for more information on the supervised ranking methods, see:

select an unsupervised classification ranking method to assess how effectively each feature performs when you do not have labeled data. the app provides two unsupervised ranking options:

selecting a method activates a new tab with a name that matches the ranking method. for more information on this tab, see ranking method tabs.

for more information on the unsupervised ranking scoring, see:

unsupervised ranking is available in diagnostic feature designer, but not in classification learner. if you plan to export your features to classification learner to train a model, you must use ensemble data that includes labels.

select a prognostic ranking method to assess how effectively each feature tracks the degradation of your ensemble members when you have run-to-failure data. the top-ranking features are best at predicting the remaining useful life (rul).

the app provides three prognostic ranking methods, all of which score features on a scale ranging from 0 through 1. one method, monotonicity, is always available. the other two methods, trendability and prognosability, are available only when you are using frame-based data. the smaller data segments in frame-based data allow the tracking of small changes that are induced by degradation.

selecting a method activates a new tab with a name that matches the ranking method. for more information on this method-activated tab, see ranking method tab.

for an example of using prognostic ranking in the app, see .

select the condition variable that provides the labels for the classification ranking algorithm to use.

specify the ranking method to sort by when comparing the results of different ranking methods. when you use a single ranking method, the app displays the results in order of importance, as indicated by the ranking score for that method. when comparing the results for multiple methods, change sort by to change the method that drives the sorting order.

specify this parameter to eliminate ranking scores for a specific method. use this parameter, for example, when comparing the results of multiple rankings, and you want to simplify the display by eliminating rankings that do not influence your feature selection.

export features to use them or share them outside of the app. both options open a ranking-sorted selectable list to choose from. when you export to the matlab workspace, you can use command-line techniques with the features. when you export to the classification learner, you open a classification learner session that uses your selected features as input.

if you want to export your entire data set from the app, use export from the feature designer tab.

you can also generate code that reproduces the computations for the variables and features you select. for more information, see the code generation options description in the export section in the feature designer tab. when you generate code using generate function for features from the feature ranking tab, ranking method defaults to the method you specify in sort by.

the correlation importance setting allows you to screen out features that convey similar information to higher ranked features. this screening provides a more diverse feature set in the upper ranks.

the criterion for the screening is the set of cross-correlation coefficients a feature has with higher ranked features. high cross-correlation between two features implies that both features are separating condition groups similarly and provide redundant information. with the default value of 0, the app does not incorporate feature redundancy into ranking scores. as you increase the correlation importance value, the app increases the influence of feature cross-correlation on the feature ranking score. this increasing influence progressively lowers the score of redundant features.

the normalization scheme performs independent normalization across the members for every feature. normalization allows more direct comparisons among features. the app displays the defining equation for the scheme you select directly beneath your selection.

this option is available only for supervised ranking and unsupervised laplacian ranking methods.

specify parameters that define key values for calculating the laplacian score, which indicates how well a given feature clusters with other features. the laplacian score is based on the pairwise distances from a given feature to its nearest neighbors.

this option is available only for the unsupervised laplacian ranking method. for more information on laplacian ranking, see fsulaplacian.

click apply to calculate a ranking with the specified parameters. the feature ranking tab in the plotting area displays the results both graphically and tabularly. this display also includes the results for the default ranking algorithm, and for any other ranking methods you calculated previously.

once you calculate a ranking, the app disables apply until you change a parameter. you can calculate a ranking within a tab multiple times. each time you modify the parameters and calculate a ranking, the new results overwrite the previous results in the plotting-area tab.

once you have completed your ranking within the ranking method tab, close that tab to return control to the feature ranking tab. the feature ranking is disabled while any ranking method tab is activated.

this parameter is read-only.

the frame policy information reflects the choice you make when you select export > generate function for... in the feature designer tab.

set criteria to refine your options when selecting items for your generated function. all criteria allow you to overwrite selectable options with a string. string matching is case insensitive. your filters apply to all output items, including signals, features, and ranking tables. criteria include:

to reset a single filter, delete the contents and click anywhere in the app. to reset all filters at once, click reset filters.

display all selected items together. use sort selected especially when you have used multiple filter combinations to assemble your codegen selections. all your selections appear together.

specify whether to use parallel computing in the generated code. the default value is the value that is specified in options. you can specify parallel computing even if you performed your interactive processing without using parallel computing. this approach helps your code to be more scalable if you plan to run the generated code on a larger ensemble than the ensemble you used to develop the features. you can also turn parallel computing off if you used it when you developed the features.

to take advantage of parallel processing in generated code, the user must have parallel computing toolbox installed and licensed. however, the code will still run in serial mode on systems that do not have the toolbox.

click the generate function button when you have completed configuring your selections. the app opens a function that contains computations used for all the output items you selected.

for more information about generating code in the app, see .