Rontala Subramaniam, Prabhakar.Vela Vela, Junior.2023-09-182023-09-1820222022https://researchspace.ukzn.ac.za/handle/10413/22277Doctoral Degree. University of KwaZulu-Natal, Pietermaritzburg.Organisations in a variety of industries are being confronted with challenging issues and trends like population changes, globalisation, and high-performance expectations. Thus, in such a competitive market, organizations have begun to pay particular attention to the recruitment and selection process, as people are their most precious assets. Employees are the most important part of any organisation as they offer values and perspectives. Employees are generally products of universities and colleges. The growth of any university depends on its ability to recruit and select qualified employees in terms of skills, knowledge, behaviour, and attitudes at all levels. However, the key aspects involved in the staff selection process, have not been thoroughly investigated. The selection process that includes interview sessions has not attracted much research. It is argued that a job interview may fail to provide a true picture of the suitability of a job candidate. As some candidates use deceptive ingratiation by claiming to correspond to the interviewers and/or organization’s values, beliefs, opinions, or attitudes to appear more appealing or pleasant, thus misleading the interviewers into selecting them for the jobs. Interview faking appears to be hard to detect, and methods for reducing it are hardly available. Nevertheless, it is important to assess the attitudinal suitability of potential academics. This can be done through various techniques such as machine learning and deep learning using social media platforms such as Twitter. Social media is an important aspect of people’s lives nowadays. As people increase their digital presence on social networking sites, the use of social media as a recruiting channel is slowly gaining momentum. This study aimed at determining how the suitability of academics can be classified using Twitter dataset. To this end, the design and development of deep learning job-fit predictive artefacts using Twitter dataset followed rigorous steps of the design science methodology. The results of this study reveal that academic suitability can be predicted using deep learning methods. This study recommends that Universities, Higher education departments consider using artefacts based on social media datasets as supplement tools to enhance the recruitment and selection process. Iqoqa Izinhlangano ezimbonini ezinhlobonhlobo zibhekene nezingqinamba nokuningi okwenzeka kule mihla njengokushintsha kwenani labantu, ezomhlaba jikelele, nokulindeleka kokwenza izinto ngokwezinga eliphezulu. Ngakho-ke, kulo mhlaba ogcwele ukuncintisana, izinhlangano seziqalile ukubhekisisa izindlela zokuqasha nokukhetha abasebenzi njengoba abantu beyifa elikhulu kuzo. Abasebenzi bayingxenye ebaluleke kakhulu yenhlangano njengoba beletha ubugugu nezindlela ezahlukene zokubuka izinto. Ngokujwayelekile, abasebenzi bawumkhiqizo wamanyuvesi namakholeji. Ukukhula kwanoma iyiphi inyuvesi kulele ekhonweni layo lokuqasha nokukhetha abasebenzi abaqeqeshiwe ngokwamakhono, ulwazi, ukwenza, kanye nendlela ababuka ngayo izinto. Nokho, izingxenye ezibalulekile ezibandakanyekayo ekukhetheni abasebenzi bezingakacwaningisiswa. Inqubo yokukhetha ngokusebenzisa inhlololwazi ibingakabhekisiswa ngendlela yocwaningo. Impikiswano iqubuka ngokuthi inhlololwazi eyimposambuzo ingahluleka ukuletha isithombe esiphelele sokuthi umuntu uwufanele ngempela yini umsebenzi. Abanye abagaqele ukuqashwa basebenzisa izindlela ezikhohlisayo ngokuphendula kahle imibuzo yenhlololwazi nokwazi kabanzi ngobugugu benhlangano, izinkolelo zayo, imibono nendlela yokwenza ukuze babukeke bethandeka ngalokho bayenge abahlolilwazi ukuba babakhethele umsebenzi. Ukukhohlisa ngezikhathi zenhlololwazi kunzima ukukubona, kanti nezindlela zokukunciphisa ziyindlala. Nakuba kunjalo, kubalulekile ukuhlola ukufaneleka kwendlela yokubuka izinto kulowo msebenzi wakwangqondonkulu ongase athathwe. Lokhu kungenziwa ngokusebenzisa amasu ahlukene njengokufunda imishini nokufunda okujulile kusetshenziswa imigudu yezokuxhumana neningi njengeTwitter. Ezokuxhumana neningi ziyingxenye ebalulekile yezimpilo zabantu ezikhathini zanamuhla. Njengoba kwanda ukuthutheleka kwabantu ezinkundleni zokuxhumana neningi, kuyakhula nokusetshenziswa kwazo njengensiza ekuqasheni abasebenzi. Lolu cwaningo lwaluhlose ukuthola ukuthi ukufaneleka kwabasebenzi bakwangqondonkulu kungahlelwa kanjani ngezigaba kusetshenziswa amaqoqo emininingo yakuTwitter. Ukuklama nokubunjwa kwezinsiza zokuqagula ukufaneleka kwalowo ozoqashwa kusetshenziswa uTwitter kwalandela izinyathelo ezinqala zendlelakusebenza yokuklama enobusayensi. Imiphumela yaveza ukuthi ukufaneleka komsebenzi wakwangqondonkulu kungaqagulwa ngokusebenzisa izindlela zokufunda ezijulile. Ucwaningo luphakamisa ukuba amanyuvesi neminyango yemfundo ephakeme iyicabangisise indaba yokusebenzisa izinto ezigcinwe yimininingo yezinkundla zokuxhumana neningi njengesu lokuthuthukisa inqubo yokuqashwa nokukhethwa kwabasebenzi.enArtificial intelligence.Deep learning.Machine learning.Social media.Twitter.Developing a predictive model using Twitter dataset for recruiting job-fit candidates in higher education institutions=Ukubunjwa kwemodeli-sifanekiso yokuqagula kusetshenziswa imininingo ekuTwitter ukuqasha abasebenzi abawufanele umsebenzi ezikhungweni zemfundo ephakeme.Thesishttps://doi.org/10.29086/10413/22277