?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=PENGEMBANGAN+SISTEM+TEMU+KEMBALI+INFORMASI+(STKI)%0D%0APUBLIKASI+DOSEN+UNIVERSITAS+LAMPUNG+MENGGUNAKAN%0D%0AMETODE+BIDIRECTIONAL+ENCODER+REPRESENTATIONS+FROM%0D%0A%0D%0ATRANSFORMERS+(BERT)&rft.creator=Adiwijaya+%2C+Satria+Nusantara&rft.subject=000+Ilmu+komputer%2C+informasi+dan+pekerjaan+umum&rft.description=Dosen+khususnya+di+Universitas+Lampung+memiliki+peran+yang+sangat+penting%0D%0Adalam+menghasilkan+pengetahuan+baru+melalui+penelitian+dan+publikasi.+Namun%2C%0D%0Aketerbatasan+sistem+yang+ada+dalam+memberikan+rekomendasi+yang+relevan+dengan%0D%0Akebutuhan+pengguna%2C+mengakibatkan+kurang+maksimalnya+penyebaran%0D%0Apengetahuan+yang+dihasilkan.+Sebagai+upaya+mengatasi+permasalahan+ini%2C%0D%0Adikembangkan+Sistem+Temu+Kembali+Informasi+(STKI)+berbasis+model+pre-trained%0D%0ABidirectional+Encoder+Representations+from+Transformers+(BERT).+Dalam%0D%0Apenelitian+ini%2C+model+BERT+digunakan+dalam+tahapan+preprocessing+untuk%0D%0Ameningkatkan+hasil+pengolahan+data+teks+mentah%2C+supaya+dapat+lebih+mudah%0D%0Adikenali+dan+dikelompokkan+oleh+sistem.+Selain+itu%2C+model+clustering+dan+juga%0D%0Asemantic+search+digunakan+untuk+membantu+mengelompokkan+teks+publikasi+serta%0D%0Amemberikan+hasil+pencarian+yang+sesuai+dengan+kebutuhan+pengguna.+Hasilnya%2C%0D%0Asistem+yang+dikembangkan+berhasil+mencapai+silhouette+score+sebesar%0D%0A0.7115418652831803+dan+memberikan+hasil+rekomendasi+yang+sesuai+dengan+kata%0D%0Akunci+pengguna+dalam+bentuk+response+web+service.+Penelitian+ini+memberikan%0D%0Agambaran+bahwa+model+pre-trained+BERT+dapat+digunakan+lebih+lanjut+dalam%0D%0ASTKI+untuk+menghasilkan+sistem+yang+lebih+akurat+dan+cepat+dalam+memberikan%0D%0Arekomendasi+hasil+publikasi.%0D%0AKata+kunci+%3A+Deep+Learning%2C+sistem+temu+kembali+informasi%2C+bert%2C+sistem%0D%0Arekomendasi%2C+web+service%2C+publikasi+ilmiah%2C+machine+learning.%0D%0ALecturers%2C+especially+at+Lampung+University%2C+play+an+important+role+in+composing%0D%0Anew+knowledge+through+research+and+publication.+However%2C+the+limitations+of+the%0D%0Aexisting+system+in+providing+publication+recommendations+based+on+users%E2%80%99+needs%0D%0Ahave+delayed+the+spread+of+knowledge+produced.+To+resolve+this+issue%2C+an%0D%0AInformation+Retrieval+System+(IRS)+was+developed+by+using+pre-trained%0D%0ABidirectional+Encoder+Representation+from+Transformers+(BERT)+model.+In+this%0D%0Aresearch%2C+BERT+model+was+utilized+in+preprocessing+stage+to+enhance+the+quality%0D%0Aof+text+processing%2C+making+it+easier+for+the+system+to+recognize+and+categorize+the%0D%0Apublication.+Additionally%2C+clustering+and+semantic+search+model+was+used+to+group%0D%0Apublication+results+and+provide+recommendation+based+on+user+needs.+The%0D%0Adeveloped+system+achieved+the+silhouette+score+of+0.7115418652831803+and%0D%0Asuccessfully+given+relevant+publications+recommendations+based+on+user+keywords%0D%0Ain+the+form+of+a+web+service+responses.+This+study+demonstrates+that+the+pre-trained%0D%0ABERT+model+can+be+effectively+utilized+in+IRS+to+create+an+efficient+and+accurate%0D%0Asystem+for+publication+recommendation.%0D%0AKeywords+%3A+deep+learning%2C+information+retrieval+system%2C+bert%2C+recommendation%0D%0Asystem%2C+web+service%2C+scientific+publication%2C+machine+learning.&rft.publisher=MATEMATIKA+DAN+ILMU+PENGETAHUAN+ALAM&rft.date=2024-06-25&rft.type=Skripsi&rft.type=NonPeerReviewed&rft.format=text&rft.identifier=http%3A%2F%2Fdigilib.unila.ac.id%2F81467%2F1%2Fopen%255B1%255D&rft.format=text&rft.identifier=http%3A%2F%2Fdigilib.unila.ac.id%2F81467%2F2%2Fopen%255B1%255D&rft.format=text&rft.identifier=http%3A%2F%2Fdigilib.unila.ac.id%2F81467%2F3%2Fopen%255B1%255D&rft.identifier=++Adiwijaya+%2C+Satria+Nusantara++(2024)+PENGEMBANGAN+SISTEM+TEMU+KEMBALI+INFORMASI+(STKI)+PUBLIKASI+DOSEN+UNIVERSITAS+LAMPUNG+MENGGUNAKAN+METODE+BIDIRECTIONAL+ENCODER+REPRESENTATIONS+FROM+TRANSFORMERS+(BERT).++MATEMATIKA+DAN+ILMU+PENGETAHUAN+ALAM%2C+UNIVERSITAS+LAMPUNG.+++++&rft.relation=http%3A%2F%2Fdigilib.unila.ac.id%2F81467%2F