搜索预测与文化遗产档案概念检测的前沿技术
1. 搜索预测算法分析
1.1 邻域算法成功率
邻域算法在不同数据集上的表现通过未加权和加权成功率来衡量,以下是详细数据:
| Set | None(未加权(加权)) | Common(未加权(加权)) | N - gram(未加权(加权)) | Fuzzy(未加权(加权)) |
| — | — | — | — | — |
| A | 4% (1%) | 5% (2%) | 9% (7%) | 10% (7%) |
| E | 54% (7%) | 55% (14%) | 58% (28%) | 59% (42%) |
| L | 14% (2%) | 13% (6%) | 19% (12%) | 27% (16%) |
| M | 73% (9%) | 81% (80%) | 39% (30%) | 45% (30%) |
| N | 5% (1%) | 8% (2%) | 11% (5%) | 16% (8%) |
| Y | 3% (1%) | 4% (2%) | 0% (0%) | 6% (4%) |
加入反馈后的邻域算法成功率有所变化:
| Set | None(未加权(加权)) | Common(未加权(加权)) | N - gram(未加权(加权)) | Fuzzy(未加权(加权)) |
| — | — | — | — | — |
| A | 4% (1%) | 12% (6%) | 14% (10%) | 15% (12%) |
| E | 54% (7%) | 61% (27%) | 69% (41