噪声数据 noisy data
Duo to its weaknesses such as great data noise, calculating complexity, inclined to overtraining, and etc.
由于其存在着数据噪声大、计算复杂、易于过训练等缺点，人们通常认为其表示能力远不如主流文本表示方法——词袋表示法。
参考来源 - 自动文本分类若干基本问题研究·2,447,543篇论文数据，部分数据来源于NoteExpress
变精度粗糙集模型允许有一定程度的错误分类率存在，利用这种特性可以有效消除投标风险信息系统中的数据噪声，提高风险要素评价的精度。
With the help of Analytical Hierarchy Process with Variable Precision Rough Set, an algorithm to evaluate significance degree of risk factors in IT project bidding was put forward.
为何不试试这种方案能否操作其自身的加密函数呢，即便数据是被加密过的，进而消除足够的噪声以使整个流程持续运转？
Why not see if the scheme could handle its own decryption function, even while keeping the data secret, to remove enough noise to let the whole process continue?
这样神经网络可应用于模式识别的特征提取、聚类分析、边缘检测、信号增强以及噪声抑制、数据压缩等各个环节。
This neural network pattern recognition can be applied to feature extraction, clustering analysis, edge detection, signal enhancement and noise suppression, data compression, such as various links.
应用推荐