Movie recommender systems are meant to give suggestions to the users based on the features they love the most. A highly performing movie recommendation will suggest movies that match the similarities with the highest degree of performance. This study conducts a systematic literature review on movie recommender systems. It highlights the filtering criteria in the recommender systems, algorithms implemented in movie recommender systems, the performance measurement criteria, the challenges in implementation, and recommendations for future research. Some of the most popular machine learning algorithms used in movie recommender systems such as K-means clustering, principal component analysis, and self-organizing maps with principal component analysis are discussed in detail. Special emphasis is given to research works performed using metaheuristic-based recommendation systems. The research aims to bring to light the advances made in developing the movie recommender systems, and what needs to be performed to reduce the current challenges in implementing the feasible solutions. The article will be helpful to researchers in the broad area of recommender systems as well as practicing data scientists involved in the implementation of such systems.

/pdf/movie-recommender-systems-concepts-methods-challenges-and-3dieu88t.pdf

Movie Recommender Systems: Concepts, Methods, Challenges, and Future Directions

Course recommendation is essential to enable learners to choose the best course among many available courses. So, as a result, a semantically driven course recommendation system is required as the data over the world wide web is increasing, and the web is also tending towards Semantic Web 3.0. In this approach, the initial term pool is created based on the existing user learner profile and the current clicks of the user to determine the exact needs of the user for the term enrichment is done by incorporating ontologies of courses and metadata generation approach is proposed where the metadata is initially classified using XGBoost algorithm where the top 25% is used. Then, the core dataset is also classified using the XGBoost algorithm based on the enriched user terms. And then, the semantic similarity is computed using concept similarity under the genetic algorithm and is recommended to the user. Overall, the accuracy of 96.3% is achieved. 

HSIL: Hybrid Semantic Infused Learning Approach for Course Recommendation

With the dawn of digital books on the web, finding reliable book suggestions has become a complicated question, and several significant machine learning algorithms were developed to address such requests. Also, there exists a problem of the prevailing recommendation system not presenting a well fit to the semantically designed nature of the world wide web. Moreover, due to the vast digitization of books, a plethora of the old book discovery strategic approaches is intended for physical selection only. Therefore, research interest is increasing in the innovative approach to search, browse, and use the recommendation engine. In this paper, an ontology focused semantic intelligence approach for the book recommendation is proposed using Jaccard similarity, NPMI, and Kullback Leibler divergence with the optimization done by the Krill Herd algorithm. The proposed system OSIBR reaches an overall Precision and Accuracy of 94.16% and 95.55% respectively and with the highest Normalized Discounted Cumulative Gain (nDCG) of 0.96. 

OSIBR: Ontology Focused Semantic Intelligence Approach for Book Recommendation

E-assessment is the method of end-to-end automated evaluation in which cyber-physical technology is used to present evaluation events and document responses. This perspective requires a full-length assessment mechanism for teachers, tutors, educational agencies, granting institutions and regulators, and the public at large. Various methodologies are implemented for increasing the quality of the assessment through automation of question generation through machine learning and deep learning concepts. Though there are various amounts of contributions toward this challenge, an efficient and a stable method has not been yet formulated. This paper introduces a process that can help to make a contribution decreasing the intensity of the challenge by using abstractive LSTM series to model sequence summarization with attention mechanism for summarization along with key generation using synonymizing with factual validation and knowledge-centric distractor generation using Wiki data for the MCQ generation. The performance of the proposed model is measured and compared to baseline models, and the proposed solution was found to be superior in terms of performance with the findings of 94.87%, 96.74%, 95.84%, and 0.03%, respectively, of precision, recall, accuracy, F-measure, and false negative rate. 

Automated Multiple-Choice Question Creation Using Synonymization and Factual Confirmation

Agriculture is a crucial source of livelihood. There are many farmers who take up farming as their main source of income. The major difficulty present among farmers is that their crop selection is not based on the soil and weather conditions. Thus, it is essential for farmers to know the necessary information regarding the different facets of crop production, best agricultural practices, suitable crops for a specific soil type to be grown in a particular weather, etc. A crop recommender system, which has accurate and precise information, serves as a means for the farmers to farm appropriate crops and get the best yield. This paper proposes an ontology-based recommender system to provide the appropriate details regarding crops and other information associated with them. The proposed OntoFusionCrop uses ontology cluster that includes soil ontology, crop ontology, geographical ontology and agricultural ontology. This strategy uses the crop recommendation dataset from Kaggle which is classified using bagging. The classified instances along with the ontology cluster are semantically aligned using spider monkey optimization algorithm from which we get facts after suitable verification. The query is asked by the user and the recommender system gives the output. The proposed model achieved 93.87% accuracy and A minimum of 0.05 FDR compared to the existing models. 

OntoFusionCrop: An Ontology Centric Approach for Crop Recommendation Based on Bagging and Semantic Alignment

Web services are software packages proposed to facilitate machine-to-machine connectivity on an interoperable network. They are the web application elements, and it is possible to post, search, and use them on the Web. The use of the web service has grown and diversified in recent years as the Internet expands. This paper offers an innovative approach for recommending web services. The intention of the user is captured by the queries input by the user searches. The search queries are collected as input and preprocessed. Query topics are generated by formalizing the input queries. The formalized queries, along with the domain knowledge, are classified using LSTM. Upon classification, the top 10% of the search results’ semantic similarity is computed using Jaccard and Lin similarity. Web service repositories such as UDDI and WSDL have been incorporated with similar terms in the dataset to recommend web services.

M. Arulmozhivarman

Papers

OWLW: Ontology Focused User Centric Architecture for Web Service Recommendation Based on LSTM and Whale Optimization