|
Compact representation of association rule with sequential item mapping rendering approach |
|---|---|
| รหัสดีโอไอ | |
| Title | Compact representation of association rule with sequential item mapping rendering approach |
| Creator | Puripat Thongkam |
| Contributor | Pakorn Leesutthipornchai, Advisor |
| Publisher | Thammasat University |
| Publication Year | 2568 |
| Keyword | Association rule mining, Boyer-Moore, Frequent pattern mining, Succinct data structures, Transactional data, การทำเหมืองกฎความสัมพันธ์, อัลกอริทึมบอยเออร์-มัวร์, การทำเหมืองรูปแบบที่เกิดบ่อย, โครงสร้างข้อมูลแบบกะทัดรัด, ข้อมูลธุรกรรม |
| Abstract | Association rule mining is widely applied to extract valuable insights from real-world datasets. However, the complex structure of transactional data makes identifying frequent item patterns computationally challenging. A key strategy to address these challenges lies in enhancing how algorithms represent item relationships. An effective representation enables efficient traversal of data structures, allowing the discovery of all embedded transactional rules. To this end, we propose Sequential Item Mapping Rendering (SIMRs), a novel framework that employs succinct data structures to optimize the representation of item relationships. SIMRs organizes these relationships hierarchically across transaction levels and compresses them into a compact string format, denoted as N1N2 ... Ni. This structured and compressed approach enhances both data processing and storage efficiency. Furthermore, SIMRs incorporates the Boyer–Moore substring search algorithm to accelerate the construction of item relationship mappings. Experimental results demonstrate that SIMRs significantly reduces memory consumption and execution time compared to both traditional and improved algorithms. Overall, the findings highlight SIMRs as a robust and efficient solution for frequent pattern mining, particularly in large-scale and complex datasets. |
