Research theme · 29 selected briefs
These topics ask how a model, dataset or analytical method behaves under a stated set of conditions. They span computing, statistics and selected applied fields where the research question genuinely uses a data method.
A software build is different from a statistical analysis: each brief identifies whether the work evaluates a model, studies data quality or compares an analytical method. Dataset access and evaluation criteria need to be settled before implementation.
Health Information Management · Data quality
This Health Information Management final year project examines how do candidate-match thresholds change duplicate flags in a synthetic patient index? The proposed evidence is synthetic identity records with known constructed matches, interpreted only for this defined research question.
Read the proposed briefArtificial Intelligence · Computer vision
This Artificial Intelligence final year project examines suggesting a limited waste category while exposing uncertain image matches through a bounded AI prototype with an explicit baseline and evaluation dataset.
Read the proposed briefData Science · Machine learning evaluation
This Data Science final year project investigates detecting when a forecasting model becomes less reliable after demand changes, with explicit data definitions and reproducible analysis rather than invented findings.
Read the proposed briefHealth Information Management · Data quality
This Health Information Management final year project examines how do blank, zero and not-recorded conventions alter summary tables in synthetic data? The proposed evidence is constructed dataset with each missingness code represented, interpreted only for this defined research question.
Read the proposed briefArtificial Intelligence · Efficient machine learning
This Artificial Intelligence final year project examines reducing inference cost while preserving a bounded intent classifier's behaviour through a bounded AI prototype with an explicit baseline and evaluation dataset.
Read the proposed briefData Science · Machine learning methodology
This Data Science final year project investigates measuring optimistic evaluation caused by selecting features before data splitting, with explicit data definitions and reproducible analysis rather than invented findings.
Read the proposed briefHealth Information Management · Data quality
This Health Information Management final year project examines which validation rules detect deliberately inconsistent values in a synthetic register? The proposed evidence is synthetic register and declared rule set, interpreted only for this defined research question.
Read the proposed briefArtificial Intelligence · Industrial computer vision
This Artificial Intelligence final year project examines detecting a small set of visible defects when labelled examples are scarce through a bounded AI prototype with an explicit baseline and evaluation dataset.
Read the proposed briefData Science · Population data quality
This Data Science final year project investigates detecting misleading rate comparisons caused by inconsistent population denominators, with explicit data definitions and reproducible analysis rather than invented findings.
Read the proposed briefHealth Information Management · Data quality
This Health Information Management final year project examines which fields show the greatest missingness in a simulated maternal-service dataset? The proposed evidence is synthetic encounter rows with declared missingness mechanisms, interpreted only for this defined research question.
Read the proposed briefArtificial Intelligence · Machine learning research
This Artificial Intelligence final year project examines measuring loss of earlier task performance after sequential model updates through a bounded AI prototype with an explicit baseline and evaluation dataset.
Read the proposed briefData Science · Predictive analytics
This Data Science final year project investigates checking whether estimated churn probabilities match observed outcomes, with explicit data definitions and reproducible analysis rather than invented findings.
Read the proposed brief