🛠 Projects
System Architecture

Aided Alzheimer’s Diagnosis System
- Designed Aided Alzheimer’s Diagnosis System based on evolving graph neural network.
- Utilized DPARSF toolkit to finish preprocessing on fMRIs from ADNI and MATLAB to construct dynamic brain network.
- Constructed E-GraphSage and E-GCN models as classifications via Pytorch geometric libs.
- Utilized ttkbootstrap libs as frontend design and Django frame as backend design.
- Completed my bachelor’s degree graduation thesis (in English) “Design and Implementation of Aided Diagnosis System of Alzheimer’s Disease Based on Evolving Graph Neural Network”.
DBMS Architecture

- Implemented a Relational Database System in C++ supporting records operations; database catalog and schema update; indexing and querying. Tested with over 100000 records/operation that finished in 60 seconds.
- Developed a paged-file manager using a Heap file structure with slot directory for efficient variable length records management. Achieved O(1) record access through optimized slot indexing.
- Developed an index manager using B+ Tree for indexing and conditional scanning on various attributes, optimizing queries, insertions, and deletions. Achieved logarithmic time complexity, resulting in a 60% reduction in query response time and a 30% improvement in data insertion speed compared to linear search methods.
- Devised a query engine that supports filter, aggregation, block/index nested-loop join and grace hash join.
System Architecture

- Designed disk external storage, partitioned disk blocks, designed file storage mode, and used Group Link algorithm to allocate free disk blocks.
- Developed comprehensive function points, including cascade create, cascade delete, copy, cut, soft link and hard link.
- Implemented the interaction based on imitative Linux query language and the parsing process of complex query functions.