Focused on the issue that the current Aspect-Based Sentiment Analysis （ABSA） models rely too much on the syntactic dependency tree with relatively sparse relationships to learn feature representations， which leads to the insufficient ability of the model to learn local information， an ABSA model fused with multi-window local information called MWGAT （combining Multi-Window local information and Graph ATtention network） was proposed. Firstly， the local contextual features were learned through the multi-window local feature learning mechanism， and the potential local information contained in the text was mined. Secondly， Graph ATtention network （GAT）， which can better understand the syntactic dependency tree， was used to learn the syntactic structure information represented by the syntactic dependency tree， and syntax-aware contextual features were generated. Finally， these two types of features representing different semantic information were fused to form the feature representation containing both the syntactic information of syntactic dependency tree and the local information， so that the sentiment polarities of aspect words were discriminated by the classifier efficiently. Three public datasets， Restaurant， Laptop， and Twitter were used for experiment. The results show that compared with the T-GCN （Type-aware Graph Convolutional Network） model combined with the syntactic dependency tree， the proposed model has the Macro-F1 score improved by 2.48%， 2.37% and 0.32% respectively. It can be seen that the proposed model can mine potential local information effectively and predict the sentiment polarities of aspect words more accurately.

Aspect-oriented Fine-grained Opinion Extraction （AFOE） extracts aspect terms and opinion terms from reviews in the form of opinion pairs or additionally extracts sentiment polarities of aspect terms on the basis of the above to form opinion triplets. Aiming at the problem of neglecting correlation between the opinion pairs and contexts， an aspect-oriented Adaptive Span Feature-Grid Tagging Scheme （ASF-GTS） model was proposed. Firstly， BERT （Bidirectional Encode Representation from Transformers） model was used to obtain the feature representation of the sentence. Then， the correlation between the opinion pair and local context was enhanced by the Adaptive Span Feature （ASF） method. Next， Opinion Pair Extraction （OPE） was transformed into a uniform grid tagging task by Grid Tagging Scheme （GTS）. Finally， the corresponding opinion pairs or opinion triplet were generated by the specific decoding strategy. Experiments were carried out on four AFOE benchmark datasets adaptive to the task of opinion tuple extraction. The results show that compared with GTS-BERT （Grid Tagging Scheme-BERT） model， the proposed model has the F1-score improved by 2.42% to 7.30% and 2.62% to 6.61% on opinion pair or opinion triplet tasks， respectively. The proposed model can effectively reserve the sentiment correlation between opinion pair and context， and extract opinion pairs and their sentiment polarities more accurately.

Some researchers suggest replacing concept lattices with concept sets in recommendation field due to the high time complexity of concept lattice construction. However， the current studies on concept sets do not consider the role of approximate concepts. Therefore， approximate concepts were introduced into recommendation application， and a genetic algorithm based Approximate Concept Generation Algorithm （ACGA） and the corresponding recommendation scheme were proposed. Firstly， the initial concept set was generated through the heuristic method. Secondly， the crossover operator was used to obtain the approximate concepts by calculating the extension intersection set of any two concepts in the initial concept set. Thirdly， the selection operator was used to select the approximate concepts meeting the conditions according to the similarity of extensions and the relevant threshold to update the concept set， and the mutation operator was adopted to adjust the approximate concepts without meeting the conditions to meet the conditions according to the user similarity. Finally， the recommendation to the target users was performed according to the neighboring users’ preferences based on the new concept set. Experimental results show that， on four datasets commonly used by recommender systems， the approximate concepts generated by ACGA algorithm can improve the recommendation effect， especially on two movie scoring datasets， compared with Probabilistic Matrix Factorization （PMF） algorithm， ACGA algorithm has the F1-score， recall and precision increased by nearly 78%， 104% and 57% respectively； and compared with K-Nearest Neighbor （KNN） algorithm， ACGA algorithm has the precision increased by nearly 12%.

As the traditional flow scheduling method for data center network is easy to cause network congestion and link load imbalance， a dynamic flow scheduling mechanism based on Differential Evolution （DE） and Ant Colony Optimization （ACO） algorithm （DE-ACO） was proposed to optimize elephant flow scheduling in data center networks. Firstly， Software Defined Network （SDN） technology was used to capture the real-time network status information and set the optimization objectives of flow scheduling. Then， DE algorithm was redefined by the optimization objectives， several available candidate paths were calculated and used as the initialized global pheromone of the ACO algorithm. Finally， the global optimal path was obtained by combining with the global network status， and the elephant flow on the congested link was rerouted. Experimental results show that compared with Equal-Cost Multi-Path routing （ECMP） algorithm and network flow scheduling algorithm of SDN data center based on ACO algorithm （ACO-SDN）， the proposed algorithm increases the average bisection bandwidth by 29.42% to 36.26% and 5% to 11.51% respectively in random communication mode， reducing the Maximum Link Utilization （MLU） of the network， and achieving better load balancing of the network.