Library · Machine learning & AI
What Feature Extraction & Selection got wrong, from 81 theses
Summary written by a model from the records below: Researchers evaluated diverse feature selection algorithms and feature extraction representations across tabular, time series, image, and text tasks. Their findings demonstrate that both aggressive feature elimination and unconstrained feature retention can harm predictive performance depending on data complexity and model type. These records come from graduate theses, 2021 to 2026, each labeled with its degree where the repository states it, and each links to its thesis. They were extracted by language models reading the full text, so treat each as a lead to read, not a verdict.
Plus 9 problems the authors left open, listed at the end.
Failure modes, by theses reporting them
Select a mode to jump to its records.
Univariate filtering and aggressive feature selection prune away critical predictive signals
Filtering variables through univariate statistics, correlation thresholds, or aggressive pruning methods consistently degraded model accuracy across multiple classification and regression problems. Preserving all candidate features or using full feature representations proved superior to discarding variables that carried subtle multivariate interactions.
Tried and failed
ANOVA feature selection applied to high-dimensional small-sample tabular data. Outcome: worse than baseline. Reason: filtering features via univariate ANOVA harmed model predictive accuracy across multiple algorithms
Predicting cancer patient response to chemotherapy using machine learning from small data · Imperial · PhD
Tried and failed
ANOVA feature selection applied to multivariate time series classification. Outcome: did not generalise. Reason: failed to yield stable or high performance across varying sequence lengths
Temporal learning techniques for vehicle cybersecurity · Texas Tech · PhD
Tried and failed
feature selection and dimensionality reduction techniques applied to FPGA compilation performance prediction. Outcome: worse than baseline. Reason: discarding features reduced the predictability performance across all models
ELECTRONIC DESIGN AUTOMATION FOR HIGH-PERFORMANCE AND RELIABLE 3D MEMORY CUBES AND PROCESSORS · Georgia Tech · PhD
Tried and failed
mRMR feature selection for dimensionality reduction applied to multivariate sensor time-series classification. Outcome: worse than baseline. Reason: aggressive feature pruning removed critical predictive signals compared to larger feature subsets or raw inputs
Evaluation of Markerless Motion Capture to Assess Physical Exposures During Material Handling Tasks · Virginia Tech · Doctorate
Tried and failed
restricting feature selection to shared platform variables applied to mass spectrometry tissue classification models. Outcome: worse than baseline. Reason: filtering for features present across both instruments discarded predictive information, lowering classification accuracy
Advancements in ambient ionization and MALDI mass spectrometry imaging methods and their application in disease diagnosis and characterization · UT Austin · PhD
Tried and failed
feature selection based on universal class correlation applied to multi-class tracking data classification. Outcome: worse than baseline. Reason: discarded discriminative class-specific features required to distinguish individual classes
Using Machine Learning to Differentiate Set Pieces in Football via Tracking Data · MIT · Master's
Lost to a baseline
Univariate feature selection underperformed PCA and full feature sets on high-dimensional genomic matrices
Applications of Machine Learning in Source Attribution and Gene Function Prediction · Virginia Tech · Master's
Lost to a baseline
Targeted feature selection had lower classification performance (lower median AUC) compared to untargeted full-spectrum models.
NMR-based metabolomic approach for the assessment of traceability, authenticity, and quality of food products · University of Trento · PhD
Lost to a baseline
Feature selection on gas-phase adiabatic ionization potential using random forest failed to improve performance and slightly worsened ionization potential predictions relative to full feature sets.
Lost to a baseline
All-features baseline beat all FS methods (KBest, mRMR, KGroups) on the TOX_171 dataset (91.15% vs 85.95% best FS).
Data Value Analytics for Feature Selection and Dataset Retrieval · University College Dublin · PhD
Lost to a baseline
Using all available features in Random Forest (RF alone) outperformed MOSS for three phenotypes, though by very small margins.
Metabolic phenotypes of marine heterotrophic bacteria · Boston University · PhD
Lost to a baseline
Monolithic reference models using all features beat all cascade variations and stacking ensemble methods in F1 score by 3.23% to 4.69% across datasets.
Promoting the perception of emerging technologies in work environments through edge computing and hybrid intelligence · University of Deusto · PhD
Considered and rejected
Considered and rejected: Rejected feature selection prior to tree-based ML training because random forest/gradient boosting ensembles intrinsically weight features and understanding every experimental condition was desired.
Advance the Synthesis of g-C3N4 Materials for H2 Production and Pollutant Removal Using Machine Learning and Response Surface Methodology · Georgia Tech · PhD
Considered and rejected
Considered and rejected: Rejected calculating ANOVA feature selection over the entire unsegmented 5-minute timecourse because full-length averaging washed out individualized functional connectivity features and produced the lowest classifier performance.
Quantifying individual differences in patterns of functional brain network organization in youths · UT Austin · PhD
Considered and rejected
Considered and rejected: Rejected feature selection during NSL-KDD preprocessing because no single subset performed equally well across all 10 evaluated classifiers.
A Detailed Analysis of the Multi-Class Classification Problem in Network Intrusion Detection using Resampling Techniques · Carleton University · Master's
Considered and rejected
Considered and rejected: Rejected selecting only the top 'n' best features for classification, choosing autoencoder feature extraction instead because models improved with more features included.
A holistic approach to improving ovarian cancer care · Georgia Tech · PhD
Tried and failed
feature selection and dimensionality reduction applied to error-related potential classification from EEG. Outcome: worse than baseline. Reason: yielded no significant performance increase over using the full feature set
Lost to a baseline
Ablating playback consistency features from the sEEG mTRF model significantly degraded prediction performance compared to the base model (EMM delta r = 0.02, p < .001, d = 0.52)
Cortical suppression of auditory feedback during speech production and perception · UT Austin · PhD
Considered and rejected
Considered and rejected: Rejected feature selection dimensionality reduction (PCA, ANOVA, Fisher Score) for EEG ErrPs because unreduced covariance and correlation features performed better.
Heuristic, wrapper, and tree-based feature selection methods overfit or exhibit ranking biases
Genetic algorithms, information criteria, and wrapper routines frequently overfit by selecting non-informative artifacts and noise when evaluating models on high-dimensional data. In addition, tree-based importance metrics like mean decrease in impurity displayed systematic biases toward continuous and high-cardinality features.
Tried and failed
genetic algorithm feature selection on concatenated multimodal data applied to high-dimensional omics classification. Outcome: overfit. Reason: selected non-biological artifacts, unannotated features, and exogenous confounders from the combined assay dataset
Metabolomics and Machine Learning for Early-Stage Cancer Diagnosis · Georgia Tech · PhD
Tried and failed
feature selection maximizing training accuracy without uncertainty constraints applied to radiogenomics regression models. Outcome: overfit. Reason: optimizing training accuracy selected too many features, increasing complexity and degrading cross-validation performance
UNCERTAINTY MITIGATION IN IMAGE-BASED MACHINE LEARNING MODELS FOR PRECISION MEDICINE · Georgia Tech · PhD
Tried and failed
heuristic threshold-based feature selection from tree importances applied to high-dimensional tabular metagenomic abundance data. Reason: empirical threshold selection retained excessive non-informative features compared to Bayesian-optimized thresholding
Antibiotic Resistance Characterization in Human Fecal and Environmental Resistomes using Metagenomics and Machine Learning · Virginia Tech · Doctorate
Tried and failed
fold-change based differential abundance ranking applied to sparse count microbiome data. Outcome: no signal. Reason: default thresholding identified very few features and truncated the rest, preventing comprehensive ranking
Statistical methods for conditional dependence and microbiome data analysis · Iowa State · PhD
Considered and rejected
Considered and rejected: Rejected standard cross-validation approaches for feature selection in small-sample microarray settings due to severe classification instability.
Power enhanced gene differential expression analysis by incorporating gene network information · Iowa State · PhD
Considered and rejected
Considered and rejected: Rejected BIC and AIC feature selection for downscaling GAMs because they selected all candidate features and overfit noise without physical justification.
Climate and chronology: environmental variability and the archaeology of marine isotope stage 3 · Oxford · PhD
Considered and rejected
Considered and rejected: Rejected simultaneously generating (X_j^+, X_j^-) for all features in Gaussian mirror due to excessive added noise and poorer feature ranking compared to the one-feature-per-fit procedure.
Problems in Variable Selection: False Discovery Rate Control and Variational Inference · Harvard · PhD
Tried and failed
Mean Decrease Gini feature importance applied to tabular random forest feature selection. Reason: ranking bias towards continuous and high-cardinality categorical features
Retrofitting U.S. Suburbia in the Era of Shared Autonomous Vehicles · Georgia Tech · PhD
Tried and failed
random forest mean decrease in impurity applied to feature importance estimation. Reason: biased toward continuous variables and features with many categories
Automated methods to find high-redshift quasars · Imperial · PhD
Tried and failed
genetic algorithm feature selection with linear regression applied to high-dimensional genomic regression datasets. Outcome: overfit. Reason: linear regression without regularization severely overfits in high-dimensional settings during wrapper feature evaluation
Application of Optimization and Simulation Models in Genomic Prediction and Genomic Selection · Iowa State · PhD
Lost to a baseline
Random Forest classifier test accuracy dropped more severely (from 0.78 to 0.69) compared to Support Vector Machine (stayed at 0.71) when pruned to top four LIME-selected features due to overfitting on the full feature set.
Multiscale Integration of Cross-Modal Subsurface Data for Reservoir Characterization under Label-Constrained Environments · Georgia Tech · PhD
Considered and rejected
Considered and rejected: Rejected global feature selection methods (Decision Trees, Random Forests, and L1-regularized Logistic Regression) because they do not capture instance-specific variability and lack runtime adaptability.
An MQTT Application-Layer Traffic Analyzer for Interpretable Flow-Level Intrusion Detection and Zero-DayThreat Identification in IoT Environment Using TabNet · York University · Master's
Handcrafted and engineered features fail to capture complex patterns compared to learned representations
Manual feature engineering such as bag of words, color histograms, and summary statistics lost temporal and contextual information required for downstream tasks. Augmenting pretrained neural representations with handcrafted metrics introduced redundancy without improving predictive performance over end-to-end models.
Tried and failed
combining hand-engineered features with machine learning models applied to competitive predictive modeling task. Outcome: worse than baseline. Reason: hybrid feature set degraded the performance relative to purely data-driven model baselines
Essays on Open Science and Open Software in Firm Innovation · Harvard · PhD
Tried and failed
Logistic regression on TF-IDF features applied to text classification of discourse tone. Outcome: worse than baseline. Reason: TF-IDF bag-of-words failed to capture subtle stylistic and rhetorical context compared to pretrained transformers
Modeling Legal Constructs · Cornell · PhD
Tried and failed
predicting hand-crafted audio features in auxiliary tasks applied to visual speech recognition front-ends. Outcome: worse than baseline. Reason: hand-crafted spectral features provided weaker representations than self-supervised learned audio representations
Deep audio-visual speech recognition · Imperial · PhD
Tried and failed
handcrafted feature extraction applied to interactive multidimensional scaling for images. Outcome: worse than baseline. Reason: color histograms and SIFT yielded less meaningful low-dimensional projections and visual explanations than CNN features
Explainable Interactive Projections for Image Data · Virginia Tech · Master's
Tried and failed
augmenting learned latent representations with handcrafted features applied to crop yield prediction. Outcome: worse than baseline. Reason: handcrafted features were redundant and highly cross-correlated with learned latent environment representations
Machine learning for mapping multimodal, multiscale phenotypic data to yield · Iowa State · PhD
Tried and failed
intermediate attribute modeling before downstream prediction applied to long-term outcome prediction from text. Outcome: worse than baseline. Reason: Intermediate human-annotated proxy features captured less predictive signal than end-to-end direct outcome modeling
EMPIRICAL EVIDENCE THAT USING AI TOOLS CAN ENHANCE HUMAN COGNITION · Penn · PhD
Tried and failed
handcrafted statistical features applied to nanopore signal classification. Outcome: worse than baseline. Reason: did not scale well with an increasing number of classes compared to recurrent neural networks
A Geometric Transformer for Structural Biology: Development and Applications of the Protein Structure Transformer · EPFL · PhD
Tried and failed
augmenting pretrained language models with handcrafted features applied to source code vulnerability classification. Outcome: worse than baseline. Reason: supplementary metrics and multi-task learning did not provide additional useful signal over pretrained representations
Deep representation learning for software vulnerability detection · Imperial · PhD
Tried and failed
direct lexical word features in discrete choice applied to social media engagement prediction. Outcome: worse than baseline. Reason: yielded lower prediction accuracy and poorer interpretability than aggregate topic features
Tried and failed
classical machine learning with hand-engineered features applied to single-molecule fluorescence time-series classification. Outcome: worse than baseline. Reason: hand-crafted timing and peak features lost critical temporal patterns needed for accurate peptide identification
Development of tools and methods for protein identification: From single molecules to in vivo applications · EPFL · PhD
Tried and failed
support vector regression with domain-specific engineered features applied to time-series signal regression. Outcome: worse than baseline. Reason: adding morphological signal-specific features degraded regression accuracy compared to standard time-series features alone
Inkjet-Printed ECG Electrodes and AI-based Data Reliability for Smart-Health CPS · Texas Tech · PhD
Considered and rejected
Considered and rejected: Rejected discarding raw sensor inputs in favour of manual feature engineering (e.g., PCA, linear discriminant analysis), selecting end-to-end multi-layer feature extraction instead
Deep learning methods for improving diabetes management tools · Imperial · PhD
Tried and failed
genetic programming for automated feature engineering applied to occupancy forecasting. Outcome: worse than baseline. Reason: Failed to provide accuracy gains over manual feature extraction
The digitalization of energy systems: towards higher energy efficiency · EPFL · PhD
Complex deep architectures and automated embeddings underperform domain-specific handcrafted features
Automated deep neural feature extractors and document embeddings were repeatedly beaten by simpler handcrafted linguistic, motion, and physical descriptors. Domain-specific features better handled signal shadowing, subjective nuance, and sparse observations while preserving model interpretability.
Lost to a baseline
On unaugmented entire-response biosensor data, SVM using TSFRESH features (F1 0.427) was beaten by SVM using 14 theory-based features (F1 0.689)
Knowledge-guided Machine Learning for Sensor-based High-Performance Autonomous Material Characterization · Virginia Tech · Doctorate
Tried and failed
raw operational parameters as ML features applied to aerodynamic flutter prediction. Outcome: worse than baseline. Reason: Derived physical and acoustic features provided stronger predictive signal than raw independent operating variables.
Data-driven modelling of compressor stall flutter · Imperial · PhD
Lost to a baseline
Doc2Vec document embeddings (59.1% SVM accuracy, 68.2% MLP accuracy) were beaten by handcrafted linguistic features (77.27% SVM, 79.09% MLP) on fake review detection
Enabling Context-Aware Natural Language Processing: From Dense Vector Representations to Contextual Features · Texas Tech · PhD
Lost to a baseline
On the CrisisLexT26SUBSET dataset, Random Forest with handcrafted features (66.80 macro F1) outperformed Random Forest with Bag-of-Words features (52.06 macro F1)
Methods for improving entity linking and exploiting social media messages across crises · Leibniz Universität Hannover · PhD
Lost to a baseline
On MCoRDS-1 Configuration I, SVM with handcrafted features achieved 99.09% OA beating Case I lightweight CNN (98.84% OA)
Novel methods for information extraction and geological product generation from radar sounder data · University of Trento · PhD
Lost to a baseline
Hand-crafted motion features outperformed Convolutional VAE and CNN-LSTM representations for predicting PTSD-Pain (AUC 0.64 vs 0.53 and 0.51).
Disease state prediction using multiscale dynamics · Georgia Tech · PhD
Lost to a baseline
Policy invocation-only features (AUC 0.776 ablated) lost to a full review content bag-of-words model, which captured subjective nuance and domain specifics better.
Leveraging Context Documents for Social Natural Language Processing · Cornell · PhD
Lost to a baseline
Automated deep feature extraction (gazePatchNet, 73.11% accuracy) was outperformed by hand-crafted feature selection using SVM (78.20% accuracy) on soccer goalkeepers.
Latent gaze information in highly dynamic decision-tasks · Tübingen · PhD
Lost to a baseline
Armeni et al. (with handcrafted features or post-processing) beat flex-convolution on the 2D-3D-S dataset for column (91.77% / 85.08% vs 52.74% AP) and door (54.11% / 45.39% vs 35.81% AP).
Efficient Processing and Learning on Unstructured Data · Tübingen · PhD
Lost to a baseline
Lang et al. outperformed the proposed MME lesion segmentation on Subject #4 (Dice 0.76 vs 0.00) due to handcrafted SNR features handling low-signal shadowing.
RETINAL OCT IMAGE ANALYSIS USING DEEP LEARNING · Johns Hopkins · PhD
Considered and rejected
Considered and rejected: Rejected raw eigenvalues alone in favor of handcrafted compound 3D features, as eigenvalues alone could not reach 3D feature performance.
Deep Semantic Segmentation of Built Heritage Point Clouds · Politecnico di Torino · PhD
Considered and rejected
Considered and rejected: Rejected using a fully deep learning-based pipeline for Heckmatt scoring in favor of PyRadiomics texture feature extraction paired with an ordinal logistic model to preserve feature explainability.
Deep learning in Ultrasound Imaging and beyond: from discriminative analysis to generative models · Politecnico di Torino · PhD
Principal component analysis and linear projections discard essential dynamics and fail on sparse or categorical data
Applying linear dimensionality reduction techniques prior to model training often degraded accuracy and obscured feature-level interpretability relative to raw inputs. Standard projection algorithms also crashed or failed when encountering zero-variance features, sparse embeddings, and low-frequency categorical variables.
Tried and failed
multi-omics factor analysis on strictly intersecting features applied to multi-tissue transcriptomics and proteomics integration. Outcome: no signal. Reason: failed to separate tissue-region variance from disease pathology variance without feature imbalance correction
Multi-omics integration in Alzheimer’s disease: linking genotype to phenotype through glial activation, synaptopathy and cognitive decline · Imperial · PhD
Tried and failed
PCA and tree-based selection for rare features applied to rare variant disease association. Outcome: no signal. Reason: Unnormalized PCA and filtering fail to capture predictive signal from extremely low-frequency categorical features.
Identifying Interaction Effects of Single Nucleotide Polymorphisms (SNPs) on Disease Risk · Texas Tech · PhD
Considered and rejected
Considered and rejected: Rejected Principal Component Analysis (PCA) because zero-variance features and categorical flag data caused standard PCA algorithms to fail.
An Analysis of Grey Theory for Personal Affective Computing · De Montfort University · PhD
Considered and rejected
Considered and rejected: Rejected including all 17 embedded categorical features in PCA/SVD due to memory/virtual machine crashes and extensive NAs in sparse port embeddings.
Intruder Alert: Dimension Reduction and Density-Based Clustering for a Cybersecurity Application · Carleton University · Master's
Tried and failed
Principal component analysis for feature extraction applied to microstructural image representations. Outcome: worse than baseline. Reason: overcomplicated data representation, degrading downstream surrogate model accuracy compared to simple summary metrics
Exploration of the Additive Manufacturing Process Development Space using High-throughput Mechanical Property Assays · Georgia Tech · PhD
Tried and failed
PCA dimensionality reduction on raw time-series applied to degradation trajectory remaining useful life prediction. Outcome: worse than baseline. Reason: linear projection discarded critical degradation dynamics captured better by domain-engineered features
Machine learning-based aging models for estimating battery state of health and predicting future degradation · Iowa State · PhD
Tried and failed
PCA dimensionality reduction before regression applied to tabular physical test data. Outcome: worse than baseline. Reason: PCA feature extraction did not systematically improve predictive performance compared to direct regression
Compacted Snow Testing Methodology and Instrumentation · Virginia Tech · Doctorate
Tried and failed
PCA dimensionality reduction before classification applied to high-dimensional tabular genomic data. Outcome: worse than baseline. Reason: Reduced classification performance compared to raw features and obstructed feature-level SHAP interpretability
Applications of Machine Learning in Source Attribution and Gene Function Prediction · Virginia Tech · Master's
Lost to a baseline
Kernel-PCA (KPCA) and Isomap feature extraction transformations were outperformed by raw untransformed power features in K-Means clustering (yielding lower Silhouette scores down to -0.15 and lower Calinski-Harabasz indices).
Study of retrofitted system for Intelligent Compaction Analyzer, a machine learning approach for Quality Control of Asphalt Pavement during Construction · University of Nevada, Reno · Master's
Lost to a baseline
PCA-based feature selection (Jaccard 0.9624) was beaten by random feature selection (Jaccard 0.985) in K-Means segmentation dimensionality reduction
MORPHOLOGICAL CLASSIFICATION OF SUBTYPES OF VOLUMETRIC PROJECTION NEURONS FROM MOUSE BRAIN SCANS · Johns Hopkins · Master's
Retaining excessive or unfiltered feature sets introduces noise and causes severe overfitting
Training complex models on raw unpruned feature spaces introduced severe noise and degraded generalization compared to compact feature subsets. Selecting a smaller set of high-quality or uncorrelated variables systematically outperformed monolithic feature concatenations.
Tried and failed
oPLS-DA on high-dimensional untargeted features without selection applied to tumor metastasis classification. Outcome: no signal. Reason: Massive feature count without prior selection introduced noise that obscured discriminative biochemical signatures
Applications of High-resolution Mass Spectrometry and Matrix-assisted Laser Desorption/Ionization Mass Spectrometry Imaging-based Non-targeted Metabolomics in Biomarker Discovery · Georgia Tech · PhD
Lost to a baseline
The addition of larger numbers of significant features resulted in lower classification accuracy in multiple hierarchical tests compared to subsets using fewer, high-quality pre-selected features.
Advanced Data Fusion Methods to Improve Wetland Classification Using Multi-Source Remotely Sensed Data · York University · PhD
Lost to a baseline
Lasso models across all SNPs in a locus underperformed top1 single-variant models in identifying predictable accessibility features (2,318 features for top1.total alone).
Allelic imbalance of chromatin accessibility in cancer identifies candidate causal risk variants and their mechanisms · Harvard · PhD
Lost to a baseline
Full multimodal feature concatenation without SHAP feature selection performed worse or barely better than unimodal models (e.g., unaware MW RF F1 was 0.057 vs unimodal Video RF F1 of 0.081).
Multimodal Machine Learning for Automated Assessment of Attention-Related Processes during Learning · Tübingen · PhD
Lost to a baseline
DNN trained with all 17 candidate features had 78.5% MAPE on unseen ramp test data, beaten by a simpler 4-parameter model which achieved 18.4% MAPE
Understanding flow-induced vibration via a physics-constrained, data-driven approach · MIT · PhD
Lost to a baseline
Deterioration prediction using univariate filter selection (F1 60% ± 2%) and no feature selection (F1 38% ± 4%) lost to Recursive Feature Elimination (RFE) wrapper selection (F1 72% ± 2%).
Wearable sensors and machine learning for tracking movement, deterioration, and recovery in acute in-patients · Imperial · PhD
Considered and rejected
Considered and rejected: Rejected relying on complex ML architectures without feature selection due to high feature-to-sample sparsity and collinearity.
PROTEOMICS IN HEART FAILURE: RISK PREDICTION AND BIOMARKER DISCOVERY · Harvard · Master's
Considered and rejected
Considered and rejected: Using all 143 raw features directly, as feature selection down to 130 features improved MAPE from 4.075% to 3.718%
The Beautiful (Computer) Game: How Data Science Will Revolutionize the World's Most Popular Sport · UT Austin
Considered and rejected
Considered and rejected: Rejected using raw full powder XRD or image pixel arrays directly as features, because p >> n in small datasets leads to severe overfitting, requiring reduction to physical descriptors.
Predicting Operational Lifetimes in Hybrid Perovskite Solar Cells: A Case Study of Machine Learning with Small Datasets · University of Washington · PhD
Considered and rejected
Considered and rejected: Rejected using all 14 handcrafted features in favor of an optimal 6-feature subset identified by Sequential Forward Feature Selection
Development of a machine vision system to estimate the physical attributes of potato tubers on-the-go at the post-harvest stage · Dalhousie · Master's
Wavelet and frequency-domain transformations introduce distortion or underperform time-domain signals
Applying wavelet transforms, spectral analysis, and frequency decompositions frequently blurred localized details or produced unstable time-frequency resolutions across bins. These transformed features failed to improve downstream regression or classification performance compared to simpler raw time-domain signals.
Tried and failed
spatial radius threshold filtering of point elevations applied to subsurface geological fault detection. Outcome: did not generalise. Reason: produced over-idealised, unvalidated straight-line linear features lacking consistent elevation offsets across points
Major faulting in London: relating tectonic-scale processes to site-scale engineering geology · Imperial · PhD
Tried and failed
directional wave component separation for feature extraction applied to pulse waveform disease classification. Outcome: worse than baseline. Reason: separated forward and backward components did not improve SVM performance over unseparated signals
Improved pulse wave analysis for diagnosing and monitoring heart failure · Imperial · Doctorate
Tried and failed
continuous wavelet transform feature extraction applied to time series correlation and regression. Reason: inconsistent time resolution across frequency bins prevented systematic correlation sweeping
Developing acoustic emission techniques for condition monitoring of rubbing contacts · Imperial · PhD
Tried and failed
wavelet transform features for neural networks applied to time-series regression. Outcome: worse than baseline. Reason: wavelet features failed with nonlinear neural models despite performing well with linear regression
Virtual metrology applied to milling process · EPFL · PhD
Tried and failed
autocorrelation shell transform denoising applied to non-smooth piecewise constant signals. Outcome: worse than baseline. Reason: standard wavelets better capture sharp discontinuities and localized non-smooth features
Topics on Multiresolution Signal Processing and Bayesian Modeling with Applications in Bioinformatics · Georgia Tech · PhD
Lost to a baseline
IQSplit feature extraction mode consistently underperformed the frequency-domain (Freq) spectral magnitude representation across both benign accuracy and noise resilience.
Device fingerprint leakage resilient on/off-body authentication and adversarial analysis of deep RF classifiers in Wireless Body Area Networks · Iowa State · Master's
Considered and rejected
Considered and rejected: Discarded the use of wavelet features as input for sliding-window decoding because it performed significantly worse across most timesteps compared to raw time-domain signals.
Decoding non-invasive brain activity with novel deep-learning approaches · Oxford · PhD
Left open by the authors
Problems the authors named and did not get to.
Left open
Perform systematic feature selection and reduction using techniques like RFE and PCA on the 17 identified search and instance features. Blocker: None
Automated design of population-based algorithms: a case study in vehicle routing · Nottingham · PhD
Left open
Benchmark co-segregation-based Bayesian linear regression feature selection against standard methods like PCA and random forests for microbial phenotypic GWAS. Blocker: None
Probing Genomic, Metabolic, And Phenotypic Evolution In Microbes Using Comparative And Experimental Evolution Data · Georgia Tech · PhD
Left open
Extract additional meta-features and evaluate a larger, more diverse pool of feature selection methods within the meta-learning framework. Blocker: No specific meta-features or feature selection algorithms are specified to evaluate
An investigation of fuzzy methods and meta learning for feature selection · Nottingham · PhD
Left open
Identify significant predictors of composite stress distributions across loading stages using feature extraction techniques on the U-Net model. Blocker: The goal is too vague regarding specific feature extraction methods and target metrics
A Deep Learning Approach to Predict Full-Field Stress Distribution in Composite Materials · Virginia Tech · Master's
Left open
Perform feature selection and incorporate hierarchical relationships between subcellular locations into CFMS protein localization classifiers. Blocker: The proposals (improve input quality, feature selection, location relationships) lack specific implementation details or defined mathematical approaches
Left open
Perform feature selection to reduce the 17 longitudinal EHR features to a minimal informative subset for delirium classification. Blocker: Requires the private clinical ICU EHR dataset used in the thesis.
Automatic delirium classification in intensive care using non-invasive eye tracking · Imperial · PhD
Left open
Evaluate the reduced feature subsets produced by the fuzzy feature selection methods using classifiers like SVM and Random Forest. Blocker: None
An investigation of fuzzy methods and meta learning for feature selection · Nottingham · PhD
Left open
Reverse engineer features learned by weak RFML learners to identify causes of recurring misclassifications without training full ensembles. Blocker: None
Sensitivity Analysis of RFML-based SEI Algorithms · Virginia Tech · Master's
Left open
Evaluate alternative feature selection methods within the class-imbalance classification pipeline to optimize feature subsets. Blocker: None
Machine Learning Approaches for Healthcare Analysis · University of Windsor · PhD
Who has worked on this
Authors of the public theses cited on this page, most-cited first.
- Sandeep Chinnareddy · Virginia Tech · Master's · 2024
Applications of Machine Learning in Source Attribution and Gene Function Prediction - Zixiao Shen · Nottingham · PhD · 2022
An investigation of fuzzy methods and meta learning for feature selection - Joseph DelPreto · MIT · PhD · 2021
Robots as Minions, Sidekicks, and Apprentices: Using Wearable Muscle, Brain, and Motion Sensors for Plug-and-Play Human-Robot Interaction - Hanqin Du · Imperial · PhD · 2026
Predicting cancer patient response to chemotherapy using machine learning from small data - Mossoun Franck Malick Jaurès Ebiele · University College Dublin · PhD · 2026
Data Value Analytics for Feature Selection and Dataset Retrieval - Anirudha Annasaheb Powadi · Iowa State · PhD · 2026
Machine learning for mapping multimodal, multiscale phenotypic data to yield - Rosamond Thalken · Cornell · PhD · 2025
Modeling Legal Constructs - Jun Wang · Georgia Tech · PhD · 2025
Retrofitting U.S. Suburbia in the Era of Shared Autonomous Vehicles - Eléonore Schneegans · Imperial · PhD · 2025
Multi-omics integration in Alzheimer’s disease: linking genotype to phenotype through glial activation, synaptopathy and cognitive decline - Arefeh KouhiRonaghi · York University · Master's · 2025
An MQTT Application-Layer Traffic Analyzer for Interpretable Flow-Level Intrusion Detection and Zero-DayThreat Identification in IoT Environment Using TabNet - Benjamin Lira Luttges · Penn · PhD · 2025
EMPIRICAL EVIDENCE THAT USING AI TOOLS CAN ENHANCE HUMAN COGNITION - Fabian Groh · Tübingen · PhD · 2025
Efficient Processing and Learning on Unstructured Data
Checking a claim in this area?
We can run the same search on any method or claim. If nothing turns up, we will say so, and that proves nothing on its own.