Skip to Main Content
Table 6.

15 potential AI solutions for one specific problem in food science and technology

 Subject (Potential solution)LevelNote
Nonlocal hierarchical dictionary learning Methods for feature selection (Zhu, Hu et al. 2016a) 
A supervised inductive manifold hashing framework Manifold hashing could be feasible for feature representation (Song, Cai et al., 2017) but may not be suitable for feature selection 
A manifold embedding algorithm Methods for feature selection (Yao, Liu et al., 2017) 
A reinforcement learning algorithm Inapplicable for this case 
Hyperspectral images Related to the problem but not a solution 
Nonparametric manifold learning Methods for feature selection (Cai, Zhang & He, 2010) 
A deterministic learning technique Methods in quantum computing, which may be theoretically applicable 
A multiobjective discrete particle swarm optimization algorithm Widely applied methods for feature selection (a large number of articles have been published in journals such as Expert Systems with Applications
A multimodal deep support vector classification (MDSVC) approach Support vector machine is one classical approach for feature selection, and the use of deep learning for extracting explicit features may be a challenge. 
10 A multikernel learning strategy Classical methods for feature selection (Zeng & Cheung, 2010) 
11 A projection-based TODIM method with MVNSs for personnel selection Classical optimization approaches, similar to #8 
12 New hashing techniques Hashing techniques are well known for data storage and transmission, but some work could be traced in the literature (Zhang, Lu et al., 2015) 
13 Existing sparse coding algorithms Methods for image/graph feature selection (Zhu, Li et al., 2016b) 
14 Spectral embedding Spectral analysis for feature selection could be traced in the literature (Li, Yang et al., 2012) 
15 A novel low-rank multiview Not a solution 
 Subject (Potential solution)LevelNote
Nonlocal hierarchical dictionary learning Methods for feature selection (Zhu, Hu et al. 2016a) 
A supervised inductive manifold hashing framework Manifold hashing could be feasible for feature representation (Song, Cai et al., 2017) but may not be suitable for feature selection 
A manifold embedding algorithm Methods for feature selection (Yao, Liu et al., 2017) 
A reinforcement learning algorithm Inapplicable for this case 
Hyperspectral images Related to the problem but not a solution 
Nonparametric manifold learning Methods for feature selection (Cai, Zhang & He, 2010) 
A deterministic learning technique Methods in quantum computing, which may be theoretically applicable 
A multiobjective discrete particle swarm optimization algorithm Widely applied methods for feature selection (a large number of articles have been published in journals such as Expert Systems with Applications
A multimodal deep support vector classification (MDSVC) approach Support vector machine is one classical approach for feature selection, and the use of deep learning for extracting explicit features may be a challenge. 
10 A multikernel learning strategy Classical methods for feature selection (Zeng & Cheung, 2010) 
11 A projection-based TODIM method with MVNSs for personnel selection Classical optimization approaches, similar to #8 
12 New hashing techniques Hashing techniques are well known for data storage and transmission, but some work could be traced in the literature (Zhang, Lu et al., 2015) 
13 Existing sparse coding algorithms Methods for image/graph feature selection (Zhu, Li et al., 2016b) 
14 Spectral embedding Spectral analysis for feature selection could be traced in the literature (Li, Yang et al., 2012) 
15 A novel low-rank multiview Not a solution 

Note that the 15 solutions were ranked based on the predicted value calculated by the proposed approach of link prediction.

Close Modal

or Create an Account

Close Modal
Close Modal