ドキュメンテーション「Audio-Based Anomaly Detection for Machine Health Monitoring​」における変数定義の​数値根拠を教えてくだ​さい。

1 view (last 30 days)
MATLABドキュメンテーション「Audio-Based Anomaly Detection for Machine Health Monitoring」におけるサポート関数
function A = getScore(data,preds)
%GETSCORE returns the reconstruction error for each sample in data
% A = getScore(data,preds) returns A(X) for each X in the set of samples
% transformed into network input data.
err = sum((preds-data).^2,2);
numSTFTFrames = 313;
contextWin = 5;
numMelFilters = 128;
numContextGroupsPerSample = numSTFTFrames-contextWin+1;
numSamples = length(err)/numContextGroupsPerSample;
A_total = reshape(err,[numContextGroupsPerSample,numSamples]); %Each column contains reconstruction errors of all context groups for one sample
A = sum(A_total)/(numMelFilters*contextWin*numSTFTFrames); %Each entry is a reconstruction error for each sample
end
における、「numSTFTFrames=313」の数値根拠(決定方法)を教えてください。
また、「contextWin=5」および「numMelFilters=128」については、一般的な数値で根拠なく任意に変更可能との認識ですが、理解は間違いないでしょか。

Answers (0)

Categories

Find more on Simulation, Tuning, and Visualization in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!