Bridging the Gap: Missing Data Imputation Methods and Their Effect on Dementia Classification Performance
Bibliographic record
Abstract
Background/Objectives: Missing data is a common challenge in neuroscience and neuroimaging studies, especially in the context of neurodegenerative disorders such as Mild Cognitive Impairment (MCI) and Alzheimer’s Disease (AD). Inadequate handling of missing values can compromise the performance and interpretability of machine learning (ML) models. This study aimed to systematically compare the impacts of five imputation methods on classification performance using multimodal data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Methods: We analyzed a dataset including clinical, cognitive, and neuroimaging features from ADNI participants diagnosed with MCI or AD. Five imputation techniques—mean, median, k-Nearest Neighbors (kNNs), Multiple Imputation by Chained Equations (MICE), and missForest (MF)—were applied. Classification tasks were performed using Random Forest (RF), Logistic Regression (LR), and Support Vector Machine (SVM). Models were trained on the imputed datasets and evaluated on a test set without missing values. The statistical significance of performance differences was assessed using McNemar’s test. Results: On the test set, MICE imputation yielded the highest accuracy for both RF (0.76) and LR (0.81), while SVM performed best with median imputation (0.81). McNemar’s test revealed significant differences between RF and both LR and SVM (p < 0.01), but not between LR and SVM. Simpler methods like mean and median performed adequately but were generally outperformed by MICE. The performance of kNNs and MF was less consistent. Conclusions: Overall, the choice of imputation method significantly affects classification accuracy. Selecting strategies tailored to both data structure and classifier is essential for robust predictive modeling in clinical neuroscience.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".