Abstract:
"Every day, a lot of social information, experience, and material are exchanged online. Since
this sharing, the quantity of user-generated material is also growing quickly, ushering in the
""age of big data."" Developers often depend on programming frameworks for the effective and
timely delivery of software. By maintaining a strong and happy user base, designers and
developers work even harder to make them accessible to their customers. These programming
frameworks, despite their best efforts, are plagued by a number of abnormalities, including
documentation, ineffective memory and compatibility-based anomalies. Q&A platform had
become one of the widespread targets for discussing this knowledge. However, the
prevalence of unorganized postings is one amongst the primary problems with the questions
and answers platforms that are now in use. This makes it challenging for the developers to
find pertinent topics. The currently in use systems were designed with little consideration for
learning state characteristics, such as natural language processing and tagged data for post
classification. Nevertheless, given the unstructured nature of the data, NLP and labelled data
alone would not be enough to detect problem postings. This research's findings have resulted
in the creation of a gap. The study suggests an improved combination system that seeks to
categorize and identify anomalies.
The project's objective is to distinguish between regular posts and those linked to anomalies
and to recommend these posts to users, designers, and developers in order to enhance
maintainability on documentation, incorrect code names, out-of-date data, and a host of other
difficulties. NLP would be used to gather the unstructured data and turn it into a structured
format. In addition, a cluster neural method would be used to increase the accuracy of the
anomaly identification.
"
