Investigation of Machine Learning on Gene Expression Data for Cancer Detection
Bibliographic record
Abstract
Cancer remains a leading cause of global mortality due to delayed diagnoses and inadequate treatments from uncontrolled cell growth.Leveraging machine learning techniques can aid in early cancer prediction given available data.This study aims to improve tumor classification accuracy and efficiency based on gene expression patterns using deep learning algorithms.The primary approach involves constructing a feedforward network (FFN) for binary classification, distinguishing between cancerous and healthy samples using the Cancer Genome Atlas (TCGA) database.Breast cancer, with ample samples in TCGA, and kidney cancer, with high mortality rates, were chosen for this study.Three feature extraction methods-Principal Component Analysis (PCA), Analysis of Variance (ANOVA), and Random Forests-were employed for preprocessing.The FFN achieved the highest accuracy for the kidney dataset using PCA with 300 principal components, yielding optimal accuracy and low error rates.For the breast dataset, PCA also produced favorable results, though requiring more principal components to retain sufficient variance.Comparative analysis showed PCA excelled in preserving variance and optimizing accuracy, with ANOVA also performing well, especially in the breast dataset, whereas Random Forests were less effective overall.These results highlight the importance of tailoring feature extraction methods and model architectures to specific dataset characteristics for the most accurate and efficient predictive models.This study demonstrates the potential of optimizing these parameters to enhance tumor classification model accuracy and reliability, providing valuable insights for improving diagnostic and treatment approaches in breast and kidney cancers.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.019 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 source (direct Gemma or distilled Codex), 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".