一种机器学习方法,用于利用电子感官特征进行人类感官快乐的预测.
Huihui Yang1,2, Yutang Wang1, Jinyong Zhao1
1Institute of Food Science and Technology, Chinese Academy of Agricultural Sciences (CAAS), Beijing, 100193, PR China.
Current research in food science
|September 11, 2023
概括
这项研究引入了一种使用融合电子感官分析和人工神经网络来预测果汁吸引力的新方法. 这项技术显示了在感官评估中取代人类感官的潜力.
科学领域:
- 食品科学与技术 食品科学与技术
- 感官科学 感官科学
- 人工智能的人工智能
背景情况:
- 人类感官评估对于评估食品质量和消费者接受度至关重要.
- 传统的感官方法可能是主观的,耗时的,昂贵的.
- 电子传感 (e-sensory) 技术提供了客观和快速的替代方案.
研究的目的:
- 开发和验证一种将融合电子感官技术与人工神经网络 (ANN) 结合在一起的方法,用于预测人类对果汁的感官快乐反应.
- 建立一个预测模型,将电子感官特征与人类感官属性和接受度联系起来.
- 探索电子感官融合作为人类感官面板的替代品的潜力.
主要方法:
- 融合电子感官分析技术被用来捕捉果汁特征.
- 人工神经网络被用来建模电子感官数据和人类感官评估之间的关系.
- 使用定量描述性分析 (QDA) 和评分测试方法进行了人类感官评估.
主要成果:
- 最初将电子感官特征与人类感官特征融合在一起的模型产生了0.77 (QDA) 和0.63 (得分测试) 的R2值.
- 最终的模型预测人类感官快乐从融合的e-感官特征实现了高准确性,与R2值为0.95 (模型-1) 和0.88 (模型-2).
- 根平均平方误差 (RMSE) 值较低,表明预测性能良好 (例如,Model-1的0.04).
结论:
- 电子传感技术与ANN的融合为预测水果汁中人类感官快乐反应提供了强大而准确的方法.
- 这种方法显示了替代或补充传统人类感官面板的巨大潜力,提供客观性和效率.
- 这项研究为开发能够同时进行多感官分析的先进设备铺平了道路.
更多相关视频
07:12Protocol for Data Collection and Analysis Applied to Automated Facial Expression Analysis Technology and Temporal Analysis for Sensory Evaluation
Published on: August 26, 2016
9.5K
09:16Multi-Modal Signals for Analyzing Pain Responses to Thermal and Electrical Stimuli
Published on: April 5, 2019
10.8K
相关概念视频
Somatosensation
36.7K
The somatosensory system relays sensory information from the skin, mucous membranes, limbs, and joints. Somatosensation is more familiarly known as the sense of touch. A typical somatosensory pathway includes three types of long neurons: primary, secondary, and tertiary. Primary neurons have cell bodies located near the spinal cord in groups of neurons called dorsal root ganglia. The sensory neurons of ganglia innervate designated areas of skin called dermatomes.
36.7K
Sensory Functions of the Skin
5.1K
The skin is the largest organ of the human body and plays a crucial role in our sensory perception. It contains a vast network of sensory receptors that contribute to the skin's protective function by perceiving physical, biological, and environmental cues and generating relevant responses.
There are two main categories of receptors on the skin: capsulated and non-capsulated. The non-capsulated ones are mainly the pain receptors. The capsulated ones can be further categorized based on the...
There are two main categories of receptors on the skin: capsulated and non-capsulated. The non-capsulated ones are mainly the pain receptors. The capsulated ones can be further categorized based on the...
5.1K
