thank you for visiting my blog, I hope my words are useful for all of you
Natural language sentiment classification algorithm based on deep learning for identifying user feelings concerning residential space
Residents can now submit subjective ratings, which are extremely valuable as subjective assessments and suggestions for architects, thanks to recent innovations in real estate brokerage systems. To analyze reviews, this work proposes a deep-learning-based natural language sentiment categorization algorithm. For 'KoNLPy' and 'Word2vec,' morphe analysis and word embedding were structured for pre-processing, and review data was processed using a long short-term memory network. This study employed a total of 5974 review data. Platforms that allow online users to give reviews of their living spaces were crawled among the different active internet platforms for real estate brokerage. The review data was labeled as 'good' or 'negative,' and the housing type was classed as 'Apartment' or 'Non-Apartment.' As additional online platforms emerge in the future and the volume of natural language data created by those platforms grows, the model developed in this study is predicted to gain in value.
Building performance evaluation (BPE) is the notion of analyzing a building's performance throughout its life cycle, from planning to design, construction, occupancy, and decommissioning (Preiser 2005). This idea encompasses not only the physical components of a building, such as its utility, legal scrutiny, and architectural requirements, but also the social, psychological, and cultural factors, as well as the people (Pexian, Thomas, and Gail 2018). Building design projects necessitate collaboration among a variety of decision-makers, including the body responsible for ordering the structure, architects, the local community, and residents. BPE plays an important role in boosting the value of the design and ensuring its quality because it is the foundation of the decision-making process at each stage of the project (Shin et al. 2017). Residents are among the many decision-making stakeholders because they will spend a significant amount of time in the final building area. As a result, incorporating a resident-centric approach into architectural design initiatives is critical. Furthermore, the physical, social, and emotional components of inhabitants' experiences in such buildings may be quantified and turned into design knowledge, which aids the architect in planning, creating, and making decisions about future projects (Goçer, Ying, and Goçer 2015).
The following is a breakdown of the paper's structure. Section 2 examines NLP, the notion of sentiment classification, sentiment classification methods, and deep learning-based sentiment classification in the context of architectural planning and design. Section 3 details the design of a deep learning-based natural language sentiment classification model after discussing methods for pre-processing natural language data and the word embedding process. The pre-processing approach is demonstrated in Section 4 along with how the findings are used to train the deep learning model. Section 5 shows scenarios in which the model can be applied, as well as a preliminary conclusion based on the results of these scenarios. Finally, Section 6 summarizes our findings and suggests some areas for future research.
This study employed the KoNLPy Korean morpheme analyser to pre-process Korean natural language data. The Word2vec model from the gensim1 Python library is used in the word embedding procedure. TensorFlow and Keras were used to create a deep learning-based sentiment classification model.
Source : https://doi.org/10.1080/00038628.2020.1748562
This article is accurate and true to the best of the author’s knowledge. Content is for informational or entertainment purposes only and does not substitute for personal counsel or professional advice in business, financial, legal, or technical matters.
© 2021 lingaji