All authors
Claude Skills by wenmin-wu
github.com/wenmin-wu535 skills0 installs688 views
- Mc Dropout UncertaintyEstimate prediction uncertainty via Monte Carlo Dropout — run inference N times with dropout active and compute mean/stdVotes: 0GitHub stars: 61
- Mean Residual DecompositionDecompose multi-output prediction into a global mean (1D model) plus per-channel residuals (2D model) with quadrature uncertaintyVotes: 0GitHub stars: 61
- Mixup Sequence AugmentationApply MixUp augmentation to padded time-series batches with Beta-distributed lambda and soft label mixingVotes: 0GitHub stars: 61
- Multi Gap Lag Diff FeaturesGenerate shift/diff features at multiple lag sizes (1,2,3,5,10,20,50,100) over cursor/time/state series, then aggregate statistics per sessionVotes: 0GitHub stars: 61
- Multi Lag Target FeaturesGenerate lag features for multiple targets over N days by shifting evaluation dates and self-joining per entity, creating a wide feature matrix of past target valuesVotes: 0GitHub stars: 61
- Multi Scale Rolling FeaturesComputes rolling mean/max/std at multiple window sizes plus total variation (abs first differences) for multi-resolution temporal context.Votes: 0GitHub stars: 61
- Multiband Color Index FeaturesCompute log-ratio features between adjacent frequency bands as color indices to characterize spectral shape from multi-band time seriesVotes: 0GitHub stars: 61
- Multiband Lomb Scargle PeriodEstimate periodicity from irregularly sampled multi-band time series using the multiband Lomb-Scargle periodogram, then phase-fold observationsVotes: 0GitHub stars: 61
- Multiband Tsfresh FftApply tsfresh per-passband feature extraction with FFT coefficients to capture multi-band periodicity from irregular time seriesVotes: 0GitHub stars: 61
- Multimodal Trajectory HeadSingle linear head that jointly predicts K candidate trajectories and K softmax confidences, sliced and reshaped for multimodal regressionVotes: 0GitHub stars: 61
- Neighbor Average Nan InterpolationFill gaps in a daily exogenous series (oil prices, sensor feeds) by merging against a full calendar to expose NaNs, then replacing each NaN with the midpoint of its nearest valid left and right neighbors, walking outward past consecutive NaN runsVotes: 0GitHub stars: 61
- Periodogram Seasonality DetectionUse scipy periodogram to identify dominant seasonal frequencies in a time series before selecting Fourier feature orders or ARIMA seasonal parametersVotes: 0GitHub stars: 61
- Poisson Lgbm Bias Corrected EnsembleTrain a single Poisson LightGBM count forecaster, then ensemble its predictions with multiple multiplicative scaling factors (alpha ≈ 1.02-1.03) to undo the systematic downward bias of Poisson regression on intermittent retail dataVotes: 0GitHub stars: 61
- Prediction Smoothing LpfApplies rolling mean or Butterworth low-pass filter to model predictions for temporal consistency and noise reduction.Votes: 0GitHub stars: 61
- Quantile Ratio ScalingConvert point forecasts to prediction intervals by scaling with logit-transformed quantile ratios passed through a Normal CDFVotes: 0GitHub stars: 61
- Quaternion Angular VelocityDerive angular velocity from consecutive quaternion frames via relative rotation and rotvec conversionVotes: 0GitHub stars: 61
- Ratio Target For SmapeTransform forecasting target to next/current ratio minus one so that optimizing MAE or squared error implicitly minimizes SMAPEVotes: 0GitHub stars: 61
- Recursive Multistep ForecastingForecast a multi-step horizon by predicting one day ahead, writing the prediction back into the panel as the new "actual", recomputing all lag and rolling features that depend on it, then predicting the next day — turns a one-step LightGBM regressor into a 28-day forecaster without changing the modelVotes: 0GitHub stars: 61
- Retroactive Outlier RescalingWalk backward through a time series and multiplicatively rescale segments when jumps exceed a fraction of the running mean to correct data collection anomaliesVotes: 0GitHub stars: 61
- Rolling Refit Arima ForecastWalk-forward validation for ARIMA by refitting on history at each step, forecasting one step ahead, then appending the true observation — produces an honest one-step error distribution that mirrors nightly-retrained production forecastersVotes: 0GitHub stars: 61
- Scaled Nan SentinelEncode missing sensor data with a per-modality sentinel value that survives standardization and remains detectable after scalingVotes: 0GitHub stars: 61
- Scaled Pinball LossScaled Pinball Loss (SPL) metric for evaluating quantile forecasts, normalized by mean absolute successive differences of training dataVotes: 0GitHub stars: 61
- Se Residual 1d Cnn1D ResNet block with Squeeze-and-Excitation channel attention for temporal sensor feature extractionVotes: 0GitHub stars: 61
- Sensor Modality DropoutRandomly zero out entire sensor modalities during training with a learned gate to handle missing modalities at inferenceVotes: 0GitHub stars: 61
- Sensor Phase VisualizationEDA visualization of multi-modal sensor data with axvspan shading for labeled behavioral phases and contiguous span detectionVotes: 0GitHub stars: 61
- Sigma Clip Outlier MaskingDetect and mask outlier data points using iterative sigma-clipping on reference frames or calibration dataVotes: 0GitHub stars: 61
- Snap Event Interaction FeaturesBuild per-state SNAP / event-flag interaction features by multiplying the binary flag with the sales and revenue columns segmented by state, capturing the demand uplift on government-benefit days that affects only specific geographies and product categoriesVotes: 0GitHub stars: 61
- Stateful Chunk InferenceProcesses long sequences in fixed-size chunks while carrying RNN hidden state across chunks for memory-efficient inference.Votes: 0GitHub stars: 61
- Store Profile Hierarchical ClusteringRe-cluster retail stores by scale-normalized weekday/dayoff mean+std profiles using Ward agglomerative clustering, replacing vendor-supplied "type/cluster" labels that correlate with store size instead of demand shapeVotes: 0GitHub stars: 61
- Temporal Frame BinningReduce temporal resolution by averaging consecutive frame blocks to improve SNR and compress high-cadence dataVotes: 0GitHub stars: 61
- Tof Spatial Region PoolingAggregate high-dimensional spatial sensor grids into hierarchical region statistics at multiple granularitiesVotes: 0GitHub stars: 61
- Transit Depth Polynomial OptimizationEstimate event depth by optimizing a scalar scaling factor on the in-event segment that minimizes polynomial baseline residual across the full signalVotes: 0GitHub stars: 61
- Tweedie Objective Zero InflatedUse LightGBM's tweedie objective with variance_power between 1.05 and 1.2 for zero-inflated count forecasting (retail SKUs, intermittent demand, click events) — handles the "many zeros plus a heavy right tail" distribution that breaks both regression (RMSE) and classification (BCE) objectivesVotes: 0GitHub stars: 61
- Unknown Class Residual ProbabilityEstimate out-of-distribution class probability as the product of (1 - p_i) across all known classes, scaled by a calibrated priorVotes: 0GitHub stars: 61
- Wavelet DenoisingDenoise an erratic 1D series with discrete wavelet decomposition + universal soft thresholding (sigma estimated from MAD of the detail coefficients) to extract the underlying trend/seasonality without lagging the signal — a far better trend extractor than rolling means for spiky retail or sensor dataVotes: 0GitHub stars: 61