Jianyang Li’s research while affiliated with Hefei University of Technology and other places

Though Deep Learning CNN is mostly used in UAV power line-inspection system for the application of intelligent image recognition technology, can design image features easily and has strong adaptability to complex environments, but three problems deafly influence the actual results of application system such as insufficient image samples library, scarce labeling samples, and absent open-data source. To conquer these problems, CBR is proposed as a strategy for knowledge reasoning, which transform the similar case-space to a new situation for problem-solving, so the combination of RBR and CBR is expected to construct our flexible case- decision diagnosis system, which integrates efficient machine learning methods to give their full advantages to guarantee the good performance of the system for fault detection. The on spot experimental results indicates our system performs efficiently, assist people in decision-making and can find potential equipment faults.

Great challenge of energy consumption in urban rail transit (URT) has been attracting much greater concerns for its more complicated impact factors. To find the energy-efficient train speed curve is the essential way for energy-saving propulsion, but it is a difficult task full of uncertainty knowledge involving in real-time train operation, which is a complicated CSP (constraint satisfaction problem) with so much inconsistent constrains that cannot be solved effectively. The paper proposes case intelligence based on CBR (case-based reasoning) to acquire the train operation preferences from the former stored cases, and constructs a flexible system integrated with efficient machine learning methods for synthesis-reasoning. The subsequent research indicates that similarity rough sets (SRS) and radial basis function network (RBF) can conquer the complexity and uncertainty of real problem, for the experimental results indicates that the hybrid system gives a fine performance as shown in real-time URT train operation.

Train energy consumption in URT has been attracted much greater concerns for it becomes more serious with the large scale operation and expansion of operation network. One of the important ways for energy-saving propulsion is to find the energy-efficient train speed curve, which is a complicated CSP (constraint satisfaction problem) with uncertainty, and cannot be solved effectively with such inconsistent constrains. The case intelligent based on CBR (case-based reasoning) is proposed in this paper for its problem-solving ability, for which the domain expertise is rich while rule knowledge deficient, to construct a flexible system integrated with efficient machine learning components and acquire the train operation preferences from the former stored cases. The experiments testing on the spot indicates that the system performs well in synthesis-reasoning, which can conquer the complexity and uncertainty of real problem from both RBR (Rule-based reasoning) and CBR, to minimize the energy consumption for train traction with punctuality and safety demands.

Attribute Weights Assignment is an important method to solve real world problems full of uncertainty, for it is difficult to acquire a comprehensive formula theoretically with which so many empirical calculations have to be drilled out of domain experts. Case-Intelligent System based on CBR (case-based reasoning), which is a human creative thinking and useful reasoning model for problem-solving, can acquire prior knowledge from the former stored cases implicating decision strategy empirical and powerful, and construct a flexible system integrated with efficient machine learning methods coping with uncertainty. Attribute weights also are the key for case similarity measurement and optimal case selection in CBR cycle, so the similarity Rough Sets is proposed for case attributes reduction, knowledge obtainment and objective weights acquiring in our case-intelligent system, which performs well in real experiments on decision-making and achieves reasonable explanations.

Though recommendation systems have been widely used for websites to generate new recommendations based on like-minded users’ preferences, IEEE Internet Computing points out that current system can not meet the real large-scale e-commerce demands, and has some weakness such as low precision and slow reaction. Huge personalized data are the key to successfully give a new recommendation, but they are difficultly dealt with for they are massive with high dimensional; addressing such problems, the paper suggests to use multi-layer feed-forward neural networks (MFNN) system based on case intelligence to partition massive personalized data into the most similar groups. The subsequent experiment indicates that our system model is constructive and understandable, and our algorithm can decrease the complexity of ANN algorithm, for which the system performance can be guaranteed.

How to suggest a valid recommend within a reasonable time is the greatest technical challenge for the recommendation system, for which tremendous user cases with high dimension are generated while it runs in real time, and these massive data are too difficult to compute directly. This paper proposes a case-intelligence system framework along with a feature-based multi-layer feed-forward neural networks (MFNN) to succeed case-retrieval based on dynamic computing, which constructs the neural networks dependence on the real input vectors instead of the fixed and dull networks structure presupposed, and can apply many kinds of knowledge granularity from various levels effectively to help users for information retrieval and case adaptation. Our subsequent experimental results indicate that it is capable of handling the massive personalized data, and our covering algorithm can decrease the complexity of MFNN algorithm for dynamic computing, which performs adaptable knowledge granularity to enhance the system's efficiency of reasoning.

Though many varieties of recommendation systems have been developed to greatly promote the intelligent level of E-commerce websites for recent years, IEEE Internet Computing points out that current system can not meet the real large-scale e-commerce demands", "and has some weakness such as low precision and slow reaction. The personalized recommendation system model based on case intelligence have proposed, which is a comprehensive expression with combination representation of human sense, logics and creativity, and can acquire the user's preferences from the former stored cases to satisfy the personalized needs. The paper focuses on how to perform effective demands on massive contents in websites, so rough sets (RS) and radial basis function network (RBF) techniques are selected to conquer problems caused by the large amounts of data. The new recommender firstly drills from the huge data in RS and reducts the main attributes, and then RBF retrieves the most valuable similar case for recommendation, which processes the same similar knowledge reasoning. The subsequent research indicates that the integrated system gives a fine performance as shown in our experiments.

Many web tools have been developed for information retrieval and information filtering to help users search, locate and manage web documents to find their needs, while IEEE Internet Computing points out that current system can not meet the real large-scale ecommerce demands, for the data from real websites is complex and multi-expression. Intelligent recommendation systems have been proposed to trace down to the acquirement of the personalized knowledge, whose system model suggests to acquire the user's preferences from the former stored cases to satisfy the personalized needs. So we select duplex techniques-rough sets (RS) and radial basis function network (RBF) - to conquer those problems caused by users' data, which are large in records, but rare in attributes. The subsequent research indicates that the newcomer can avoid sensitive to the noise along with the influence of irrelevant items, with which the results of our experiments for the validation on the proposed model are applausive.