48 research outputs found
Pengaruh Pemberian Mikoriza dan Kompos Tricoderma sp (trichokompos) Terhadap Pertumbuhan Vegetatif Tanaman Glycine max (l.) Merill)
Full text linkKedelai merupakan komoditas pangan utama setelah beras dan jagung dengan
kandungan protein nabati yang tinggi. Rendahnya produksi kedelai di Indonesia,
salah satunya dipengaruhi oleh penurunan kesuburan tanah yang diakibatkan
adanya penggunaan pupuk anorganik secara berlebihan. Sehingga penggunaan
mikoriza dan trichokompos diharapkan mampu memperbaiki struktur tanah dan
penyerapan unsur hara tanaman. Tujuan penelitian ini adalah untuk mengetahui
pengaruh pemberian mikoriza, trichokompos serta kombinasi mikoriza dan
trichokompos terhadap pertumbuhan tanaman kedelai (Glycine max (L.) Merill).
Penelitian dilaksanakan pada bulan Juli sampai September 2024, di Desa
Tuntungan II, Kec. Pancur Batu, Kabupaten Deli Serdang, Provinsi Sumatera Utara.
Penelitian ini merupakan penelitian eksperimen dengan menggunakan metode
Rancangan Acak Kelompok (RAK) faktorial dengan 2 faktor dan 3 ulangan. Faktor
pertama Mikoriza dengan 3 taraf : tanpa mikoriza (M0), mikoriza 20 gram (M1),
dan mikoriza 40 gram (M2) dan faktor kedua trichokompos yaitu tanpa
trichokompos (P0), trichokompos 4 kg (P1), dan trichokompos 8 kg (P2). Data yang
diperoleh dianalisis menggunakan sidik ragam dan dilanjutkan uji jarak berganda
duncan (DMRT). Hasil penelitian menunjukkan bahwa pengaruh pemberian
mikoriza berpengaruh nyata terhadap tinggi tanaman, total luas daun, dan laju
pertumbuhan relatif. Pemberian trichokompos berpengaruh nyata terhadap tinggi
tanaman, total luas daun, dan laju pertumbuhan relatif. Pemberian kombinasi
mikoriza dan trichokompos berpengaruh nyata terhadap total luas daun dan laju
pertumbuhan relatif
ANALISIS STRUKTUR PENGENDALIAN INTERN ATAS PENERIMAAN KAS DARI PENJUALAN TUNAI PADA PT SINAR BEARINDO SENTOSA PALEMBANG
Full text linkThe Internal Control Structure Analysis On Cash Receipts of Cash Sale in Sinar Bearindo Sentosa, Ltd. Palembang
Mutia Silvi Aprialina, 2015 (xii + 59 pages)
E-mail: [email protected]
The author conducted data collection by means of observation and interviews directly to the object of research. Analysis was conducted by the author in completing this final report is in the form of an analysis of the internal control structure on cash receipts of cash sale that have been implemented enterprise. From the results of the analysis, the author found several problems which were geminating task by the cashier function, cash and credit sales invoices were not separated, did not have work guidelines, did not have a policy of hiring, and absence of checks cash receipts transactions. After analyzing the existing problems, the author gives some suggestions that can help the development of the company in the future are companies preferably separates accounting functions and cash and have a policy in reception clerks, Sinar Bearindo Sentosa, Ltd. Palembang should make cash and credit sales invoices separately, and preferably Sinar Sentosa Bearindo, Ltd. Palembang has guidelines for employees, and cash receipts transactions examination
Semiotics as science
Full text linkThe present article gives an overview of different approaches on semiotics as science, its objects of investigation, methods and genesis (where, how and when does semiotics begin?). The author does not aim at establishing one prescriptive approach. Quite the opposite, by leaving the question open, the author aspires to encourage further discussion about the criteria for scientificity, establishing the borders of scientific disciplines, and the productivity of the dialogic (or, rather, polylogic) scientific meta-discourse in science in general and in semiotics in particular
Costume development for Jõhvi`s Folk Dance Group "Gevi"
No full textDiplomitöö käigus töötati välja Jõhvi rahvatantsurühma „Gevi“ kostüümi osad, mille kampsunite aluseks on Eesti Rahva Muuseumi kogudest pärit Simuna kampsunid ja kleidi inspiratsiooniks tellija poolt esitatud pilt. Töö käigus lähtuti kliendi soovidest; individuaaltöö ning kostüümi- ja rahvarõivaste valmistamise eripäradest. Autor konstrueeris toodete lõiked ja töötas välja tehnoloogia. Töö tulemusena õmbles autor valmis ühe rahvatantsija villase- ja ruudulise kampsuni ning kleidi suuruses 42. Ülejäänud tantsijate kleidid õmmeldakse rahvatantsuseltsi poolt palgatud õmbleja poolt. Lekaalikomplektid valmisid standardsuuruste tabeli mõõtude järgi ja kohandatakse proovide käigus iga tantsija personaalsete mõõtude järgi. Lekaalide põhjal on edaspidi võimalik unifitseerida ka teiste kihelkondade kampsuneid. Kostüümi kampsunite tehnilise töötlemise viis erineb oluliselt tänapäevasest, mistõttu töödeldi paljud lõikeservad ja ääred puhtaks käsitsi. Tantsijate liikumisvabaduse saavutamiseks lisati varrukatele kaenlalapid, mis võimaldavad paremini käsi tõsta. Lisaks pidi tehniliste lahenduste loomisel arvestama tantsijate vahetumisega rühmas, mille tõttu pidid tooted olema korrigeeritavad. Diplomitöö jagunes seitsmeks peatükiks, mis on järjestatud lähtudes individuaaltööna valminud toodete valmistamisest. Töös kirjeldatakse toodete ajaloolist tausta ja inspiratsiooni, lekaalide valmistamise põhimõtteid, põhi- ja moekohaste lõigete valmistamist M. Müller & Sohn süsteemist lähtudes ning unifitseerides, lekaalide paljundamist suurustele 38-52, materjalide ja furnituuride valikut, kangakulu arvestust ning tehnoloogiat. Seitsmendas peatükis kirjeldatakse tootmiskulude arvutamist, võttes aluseks ettevõttes juurutatud meetodi. Töös kajastuvad joonised on tehtud järgmiste programmidega: Optitex, Optitex Marker ja Corel DESIGNER. Autori jaoks oli diplomitöö koostamine huvitav ja uudne kogemus. Töö eelduseks oli Eesti Rahva Muuseumis originaalesemete uurimine ja ajaloolise materjaliga tutvumine, mis tagasid põhjalikud teadmised tehtavast. Kuigi autentsuse tagamiseks tuli välja mõelda erinevaid tehnoloogilisi lahendusi, on autor tulemusega rahul. Suur abi oli diplomitöö ettevõttepoolse juhendaja teadmistest ja kogemustest, mille tõttu jäid paljud vead tegemata. Kuigi valminud lekaalide juures on aspekte, mida saab edasise arendamisega parandada, võib esimesed tooted lugeda õnnestunuks.The aim of the final thesis was to create linen dress, woolen and checked long-sleeved jacket for folk dancers group. Jackets had to be based on Simuna`s long-sleeved jackets from the collections of the Estonian National Museum. The dress was inspired of picture, which was given by customer and had to fit to all of the dancers. The author was tasked with drafting the patterns, creating technical solutions and tailoring costumes. Jõhvi`s woman`s group Gevi has been engaged with folk dance in Ida-Virumaa for 53 years. In 2001. founded non profit organization Jõhvi`s Folk Dance Association Gevi, to manage and lead group`s activities. Association`s main goal is to develope and preserve Estonian national culture. [1] Silvi Allimann, the founder of Kaunis Rahvarõivas Oü, has been engaged making folk costumes for 40 years [2]. Earlier, her company has made folk costume set for „Gevi“. Set included woolen skirt with longitudinal stripes, embroidered blouse, cap, apron, undershirt, belt, long-coat and kerchief. Folk dance costumes are different from regular garments. Their seam allowances must be longer, because costumes are fitted to each dancer and many cutting edges are processed manually. Also there is added armpit patch to give dancers freedom for movement. This graduation thesis consists of seven chapters, each of them describes different part of costume development. The first chapter describes historical long-sleeved jackets and explains why Simuna`s jackets were chosen. Also there is pointed out the changes, which were made based on dancers needs. The chapter concludes description of dress designing and clients demands for the item. The following part describes models, their details, materials and attachment. There are also technical drawings and explanations for the choice of the materials. The third chapter contains the base bodice pattern drafting according to M. Müller & Sohn method. There are also described modifications which are made for patterns and measurements which are used for constructions. The fourth paragraph covers the technical solutions and technical sectional drawings. Resolutions are described in table of technical processing. The following part reviews the development of master patterns, seam allowances and grading rules. The sixth chapter describes lay planning and calculations in material usage effiency. There are compared combined and consolidated layout. The seventh paragraph contains calculation of producton costs using the company`s method. Working on this thesis has been interesting and fresh experience. The assumption of this work was investigation of original objects in Estonian National Museum and study of historical material. These methods gave good basic knowledge. Although different technological solutions had to be made to ensure authenticity, the end result is satisfying. Company`s tutor`s knowledges and experiences were a great help in making this thesis and though master patterns have aspects which can be improved during further work, the first products can be considered successful
Petite histoire d’une appropriation réussie : François Buffereau plagiaire de Gossuin de Metz
No full textThe Short Story of a Successfull Arrogation : François Buffereau, Plagiarist of Gossuin de Metz.
The only known and ancient manuscript of the first version in verse of Gossuin de Metz’s Image du monde was published in 1517 by Jacques Vivian, a printer in Geneva. It is in fact a case of plagiarism as a certain François Buffereau, devoted secretary of Antoine de Gingins, Lord of Divonne, introduces himself as the author of this work in the prologue and the explicit. This study aims at determining what Buffereau personally adds to this medieval encyclopedia, the contents of which are outdated and, if his arrogation seems to be dishonest, at ascertaining whether his work succeeds in any way. At first, in order to substantiate his undertaking, the plagiarist does not hesitate to stage his reading and writing works. Then, he changes slightly the very matter of the treatise : if the true nature of his interventions is quite vague for the most part – is he responsible for the modernization of the language, or is it due to the copyist before him or even to the printer after him ?– his liability is engaged when it comes to the paratext, especially in the margins that are filled with bibliographical references made by the little ducal secretary, which are definitely Buffereau’s personal contribution. As they change the Image into a Mirror, these marginal indications allow our plagiarist to accede to the much coveted authorship.C’est en 1517 que paraît chez Jacques Vivian, imprimeur à Genève, la seule édition ancienne connue à ce jour de la première rédaction en vers de l’Image du monde de Gossuin de Metz. Il s’agit en fait d’un plagiat, un certain François Buffereau, secrétaire dévoué d’Antoine de Gingins, seigneur de Divonne, se présentant, dans le prologue et dans l’explicit, comme l’auteur de l’ouvrage. L’objectif de cette étude est d’essayer de déterminer quel est l’apport personnel de Buffereau à cette encyclopédie médiévale au contenu suranné et, si l’appropriation peut sembler malhonnête, de voir si elle est au moins réussie. D’abord, et pour accréditer son entreprise, le plagiaire n’hésite pas à mettre en scène son travail de lecture et d’écriture. Ensuite, il fait subir quelques changements à la matière même du traité : si la véritable nature de ses interventions sur le texte reste, à quelques exceptions près, assez floue – est-ce lui ou un copiste avant lui, voire l’imprimeur après lui, qui l’a raccourci et qui en a modernisé la langue ?–, sa responsabilité est en revanche engagée lorsqu’il s’agit du paratexte, et plus précisément de celui qui se situe dans les marges du texte, que le petit secrétaire ducal remplit de références bibliographiques, dont nous avons essayé d’étudier les modalités et qui constituent, à coup sûr, l’apport personnel de Buffereau. Ce sont ces indications marginales qui, transformant l’Image en Miroir, permettent à notre plagiaire d’accéder au statut tant convoité d’auteur.Christine SILVI, Verhaal van een oneerlijke toe-eigening : François Buffereau, plagiator van Gossuin de Metz.
In 1517 verscheen bij Jacques Vivian, drukker in Genève, de enige oude editie, voor zover bekend, van de eerste verzen op schrift van Image du monde van Gossuin de Metz. Het betreft hier eigenlijk plagiaat, een zekere François Buffereau, een toegewijde secretaris van Antoine de Gingins, seigneur van Divonne, presenteert zich in de proloog en in het explicit, als de auteur van het werk. Het doel van dit onderzoek is om aan te tonen wat de persoonlijke inbreng is geweest van Buffereau bij deze middeleeuwse encyclopedie met zijn verouderde inhoud. Daarnaast om te bekijken of deze succesvol is geweest, ook al was de toe-eigening niet eerlijk. Allereerst zal de plagiator om de geloofwaardigheid van zijn werkzaamheden te vergroten zoveel mogelijk doen alsof hij zich bezighoudt met lezen en schrijven. Vervolgens brengt hij enkele wijzigingen aan in de inhoud van de verhandeling : als het soort wijzigingen in de tekst, op enkele uitzonderingen na, vaag genoeg blijft, was het dan de plagiator of de kopiïst voor hem, of zelfs de drukker na hem, die hem heeft ingekort en het taalgebruik heeft aangepast ? Daarentegen kan de plagiator wel verantwoordelijk gehouden worden als het gaat om paratekst, en in het bijzonder om tekst die zich in de kantlijn bevindt. De kantlijnen bevatten vele bibliografische referenties van de hand van de hertogelijke secretaris, waarvan wij de modaliteiten hebben getracht te bestuderen, en die zonder enige twijfel behoren tot de persoonlijke inbreng van Buffereau. De marginale aanwijzingen die het Beeld (Image) veranderen in een Spiegel (Miroir) bieden onze plagiator de mogelijkheid om de begeerde status van auteur te bemachtigen.Silvi Christine. Petite histoire d’une appropriation réussie : François Buffereau plagiaire de Gossuin de Metz. In: Revue belge de philologie et d'histoire, tome 94, fasc. 3, 2016. Langues et littératures modernes – Moderne Taal- en Letterkunde. pp. 661-689
Analisis Perbandingan Metode Bandpass Filter dan Bandreject Filter pada Domain Frekuensi untuk Mereduksi Noise pada Citra Digital
No full textNoise adalah gangguan pada citra digital yang dapat menghilangkan infromasi yang dibutuhkan pada citra tersebut. Salah satu mekanisme yang digunakan untuk mengurangi noise adalah filter. Adapun metode filter yang digunakan oleh penulis umtuk mengurangi noise adalah metode Bandpass Filter dan Bandreject Filter yang dilakukan dalam domain frekuensi. Oleh karena itu, penulis juga menggunakan metode Fast Fourier Transform untuk menghasilkan citra transformasi di dalam domain frekuensi. Implementasi sistem menggunakan IDE SharpDevelop 4.3. Hasil penelitian menunjukkan bahwa citra hasil filter menggunakan Bandpass Filter lebih baik kualitasnya dibandingkan dengan citra hasil filter dengan Bandreject Filter. Rata – rata MSE dan PSNR yang dihasilkan metode Bandpass Filter adalah 24209,191 dan 4,307 dB. Sedangkan rata – rata yang dihasilkan metode Bandreject Filter adalah 2137,811 dan 14,704 dB. Dari segi waktu dapat diketahui bahwa semakin besar ukuran citra yang diuji semakin besar pula waktu yang dibutuhkan dalam pelaksanaan proses filter. Jika dilihat dari tampilan visual, Bandreject Filter memberikan hasil citra yang lebih baik dibandingkan dengan menggunakan Bandpass Filter.Noise is an interferences in digital image that can eliminate the required information on the image. One of the mechanism used to reduce noise is filtering. The filter method used by the author to reduce noise is Bandpass Filter and Bandreject Filter which are both performed in the frequency domain. Therefore, the author also use Fast Fourier Transform method to generate image transformation in the frequency domain. System is implemented by using SharpDevelop 4.3 IDE. The results showed that the image has filtered by using Bandpass Filter better quality than image has filtered by using Bandreject Filter. The average of MSE and PSNR produced by Bandpass Filter method is 24209,191 and 4,307 dB. While the average result of Band Reject Filter method is 2137,811 and 14,845 dB. In terms of time can be seen that the larger size of the image that tested the greater time required in the filter process. When viewed from the visual display, Bandreject Filter provides better image results than using Bandpass Filter.128 HalamanSkripsi Sarjan
Improvement of the Production Process of Men’s Blazers and Women’s Jackets
No full textAutor andis töös ülevaate ettevõttes Baltika Tailor OÜ toimuvast ja tegi ettepanekuid, kuidas lahendada probleeme ja selle abil parendada tootmisprotsessi Lõputöös on uuringud tehtud pintsakuliini kahe esimese sektsiooni põhjal- Pintsak 1 ja Pintsak 2. Vaatluse alla on võetud meeste pintsak „Asoft 5”- 207 toodet ja naiste jakk „Pabou”-300 toodet. Töökäigus on välja toodud toodete töötlemisjärjekord, protsessitüüp, tööjaotus, graafiline analüüs, balansseerimine ja päevavaatlused. Kuna ettevõtte sooviks oli tootmisprotsessi parendamine, sai algselt koostatud koondoperatsioonide tabel, mille eesmärgiks oli välja tuua töötajate ala- või ülekoormatus. Pintsakute liinis olevad tooted on tihti väga erineva keerukusega ja töötlemise ajad väga erinevad. Tööjaotus koostati selle järgi, nagu on see ettevõttes. Tööjaotamisel oli arvestatud ka töötajate efektiivsust. Ettevõttes on enamus töötajatest 100 % ja rohkema efektiivsusega. Eelbalansseerimine sai koostatud selleks, et operaatorite töö sisu sobituks paremini nende oskustega ning tagaksrütmilisema töö liikumise. Seejärel tehti reaalne liini balansseerimine, mis andis ülevaate liinis toimuvast. Selleks, et tööd saaks jaotada võimalikult paindlikult, on vajalikud tööliste erinevad oskused. Hetkel on küll töötajatel palju erinevaid oskusi, kuid kardetakse töötamist spetsmasinatel ja automaatidel, sest seal on vastutus suurem. Seega julgustaks töötajaid juurde õppima uusi keerulisemaid operatsioone sektsioonisiseselt, siis ontööjuhil ka lihtsam liini balansseerida.Omalt poolt soovitaks ettevõttele luua ka uus R-kategooria, mille eesmärgiks oleks õpetada juurde operatsioone ka liini teistest sektsioonidest.Uue kategooriaga tuleks arvestada ka tükitöö hinde tõstmisel. Lisaks sai balansseerimisel välja uuritud, milline töösse laskmise viis on kõige mõistlikum. Sellega sai ettevõtte kinnitust, et praegune tootmisse suunamine - paralleelselt meeste pintsak ja naiste jakk - on kõige õigem. Paraku alati pole võimalik seda teha. Võimalusel võiks lasta sel juhul paralleelselt töösse kaks naiste jakki - üks keerukam ja kõrvale kergem naiste jakk. Lõputöö käigus tehti töötajatele päevavaatlust, mis koostati kahele õmblejale ja ühele tööjuhile. Õmblejate päevavaatlusest selgus, et töötajad on väga erineva tööaja kasutusega. Päevavaatlus viidi läbi kahele sarnase operatsiooniga õmblejale. Mõlemad töölised õmblevad taskuid - üks põhimaterjalist taskuid ja teine voodritaskuid. Uuringu käigus oli näha, et töötajad tulevad tööle erinevatel aegadel, kuid see ei sea takistusi tootmisel. Töötaja teab, et peab suutma tagada järgnevale töötajalepiisava pooltoodangu varu järgmiseks päevaks. Kindlasti soovitaks töötajatele, kes ei käi ettenähtud pausidel, teha kindlasti pause, kuna peale väikest puhkust on töötaja efektiivsem. Samuti võiks töötajatele rohkem meelde tuletada, et nad teeks ka väikesi paariminutilisi võimlemis- ja venitusharjutusi. Tööjuhi päevavaatlusel oli koheselt näha, et tal on palju tööd. Sel päeval, kui vaatlus läbi viidi, tuli ette praagi tagastust.Praagi eemaldamiseks kulus tööjuhil antud vaatluse ajal liialt aega. Parem oleks, kui tööjuht saaks seda aega suunata töötajate õpetamisele ja instrueerimisele. Ka vaatluspäeval oleks olnud otstarbekas praagi edaspidisele vältimisele suunatud töötajate juhendamine. Kõrgendatud tähelepanu kvaliteedi jälgimisele tuleks rakendada eriti siis, kui töösse tuleb uus mudel.Tööjuhid võiksid töötajaid rohkem usaldada kaasates neid probleemide lahendamisele. Kvaliteedijuhtimine on üks väga oluline osa tootmisprotsessis. Selleks et olla jätkusuutlik ja konkurentsivõimeline, peab ettevõte olema pidevas parendamise protsessis. Baltika Tailori tootmises on põimunud erinevad kvaliteedijuhtimise süsteemid - Sixsigma, LeanProduction ja Just-in-time. Kvaliteedijuhtimise veelgi paremaks juhtimiseks pakuks omalt poolt välja DeNisi ja Klugeri süsteemi, mis tugineb kolmele põhimõttele: isikule suunatud, ülesandele suunatud ja ülesande õppimisele suunatud tagasiside.Hea kommunikatioon tagab parema kvaliteedi, rohkem oskusi ja rütmilisema töö. Lõputöö autorina soovin tänada Baltika Tailor OÜ pintsakute liini vanem tööjuht-tehnoloogi Silvi Puhit ja Tallinna Tehnikakõrgkooli poolset juhendajat Anu Tšistovat.The following this thesis is Improvement of the Production Process of Men’s Blazers and Women’s Jackets. The aim of the research was to improve the production process of the blazer production line in Baltika Tailor OÜ. The author provides an overview of the current state of the company’s production process as well as suggestions for its improvement by addressing existing issues. The research used in the thesis is based on the first two sections of the blazer production line, Blazer 1 and Blazer 2. The products surveyed were the men’s suit blazer Asoft 5 (207 items) and the women’s jacket Pabou (300 items). The following facets of the research are presented: the product processing sequence; the processing type; the division of labour; graphic analysis; balancing; and observation of the working day. Because the company wished to improve its production process, initially a table of aggregated processes was compiled. The aim of this was to highlight the insufficient or excessive workload that employees may experience. Products in the blazer line often vary considerably in complexity, and process durations are very different. The division of labour drawn up followed that of the company. In dividing up the labour, the efficiency of the employees was also considered. The majority of employees in the company were shown to be at least 100% efficient. Pre-balancing was conducted in order to match operators’ tasks to the operators’ skills and thus to ensure the more rhythmic movement of the process. After this, true production line balancing was conducted. This produced an overview of the work on the line. In order to divide up labour as flexibly as possible, employees need different skills. At the moment the employees have many different skills, but there is a reticence to operate specialised machinery and automatic equipment. This is because operating such machinery and equipment comes with greater responsibility. Therefore I would recommend that employees be encouraged to learn new and more complex operations. This would make it easier for the production supervisor to ensure balance on the production line. Another aspect of balancing was considering methods of the organisation of production for efficiency. This reassured the company that their current method of organisation is the most efficient. Said method is to begin work on a men’s blazer and women’s jacket simultaneously, but this is not always possible. In such a case, working on two women’s jackets at the same time should be considered, if possible. One of these should be more complex and the other more simple to make. Observation of the working day of the employees was also conducted as part of the research. Two tailors and one production supervisor were observed. Observation of the tailors showed that the employees make different use of their working hours. Observation was conducted on two tailors with similar tasks. Both employees sew pockets: one with the main material and the other with the lining. The study also showed that employees start work at different times. This, however, does not inhibit production. The employees know they must ensure an adequate stock of semi-finished products for the next day and the next employee. I would definitely recommend that workers who do not make use of their designated breaks do so. Taking small breaks would certainly increase their productivity. I also recommend reminding employees to take a few minutes to do stretching exercises and callisthenics. The observation immediately showed that the production supervisor has a lot of work. On the day of the observation there were cases of defective product returns. During the observation the supervisor spent a little too much time handling these products. It would be better if the supervisor used this time to instruct and advise employees. Attention to quality could be increased, particularly when a new model is introduced to production. Quality management is a crucial part of the production process. In order to be sustainable and competitive, a company must constantly improve. Baltika Tailor has incorporated several quality management systems into its production, these being Six Sigma, Lean Production and Just-In-Time. To further improve quality management I would recommend a system based on ISO standards. This would present requirements for quality management, but would not dictate their method of implementation within the company
Sistem Analisis Sentimen pada Fanpage Facebook Kandidat Presiden 2019-2024
No full textFanpage kandidat presiden merupakan media untuk berkomunikasi dan bertukar informasi dengan simpatisannya. Selain itu, fanpage juga seharusnya berisi lebih banyak sentimen positif dengan asumsi bahwa pengikut halaman adalah orang yang mendukung. Namun ada beberapa oposisi yang tidak sependapat sehingga timbul sentimen negatif. Dalam penelitian ini, penulis akan melakukan analisis sentimen berdasarkan komentar yang terdapat dalam fanpage kandidat presiden dengan menerapkan proses text mining dan algoritma Support Vector Machine (SVM) dan K-Nearest Neighbours (K-NN) untuk mengklasifikasikan apakah teks termasuk dalam sentimen positif, negatif dan netral. Penulis menggunakan kurang lebih 6000 komentar dari masing-masing fanpage kandidat presiden sebagai data latih yang akan diklasifikasikan sentimennya. Data diklasifikasikan secara manual dengan mengelompokkan menjadi kelas sentimen postif, negatif dan netral kemudian secara otomatis data latih akan mengambil beberapa data untuk data uji dan menguji kemampuan sistem. Hasil penelitian berupa persentase akurasi menggunakan 2 algoritma. Algoritma K-Nearest Neighbours terbukti mendapatkan nilai akurasi lebih tinggi dari algoritma Support Vector Machine yaitu pada fanpage Joko Widodo dan Ma’ruf Amin akurasi yang diperoleh algoritma K-Nearest Neighbours sebesar 94 % sedangkan algoritma Support Vector Machine sebesar 47%. Pada fanpage Prabowo Subianto dan Sandiaga Uno akurasi yang diperoleh algoritma K-Nearest Neighbours sebesar 96% sedangkan algoritma Support Vector Machine sebesar 65%.The presidential candidate's fanpage is a medium for communicating and exchanging information with sympathizers. In addition, the fanpage should also contain more positive sentiment assuming that the followers of the page are people who support. But there are some opposition that disagrees so that negative sentiment arises. In this study, the author will conduct sentiment analysis based on comments contained in the presidential candidate fanpage by applying the text mining process and the Support Vector Machine (SVM) algorithm and K-Nearest Neighbors (K-NN) to classify whether the text is included in positive, negative sentiments and neutral. The author uses approximately 6000 comments from each presidential candidate fanpage as training data to be classified sentiment. Data is classified manually by grouping into positive, negative and neutral sentiment classes and then automatically training data will take some data for test data and test the ability of the system. The results of this research are percentage accuracy using 2 algorithms. The K-Nearest Neighbors algorithm is proven to get higher accuracy values than the Support Vector Machine algorithm, which is on the Joko Widodo and Ma'ruf Amin fanpage, the accuracy obtained by the K-Nearest Neighbors algorithm is 94% while the Support Vector Machine algorithm is 47%. In the fanpage of Prabowo Subianto and Sandiaga Uno the accuracy obtained by the K-Nearest Neighbours algorithm is 96% while the Support Vector Machine algorithm is 65%.112 HalamanSkripsi Sarjan
Question Answering System for Climate Crisis Topic with Sentence-BERT Model
No full textIn obtaining information among the many documents with various topics stored on the internet, it is a challenge for users to choose the right information from each search result. This is one of the triggers for users to be exposed to disinformation on important topics such as the climate crisis, which will have an impact on hampering climate change mitigation efforts. In this research, the author builds a question answering system that can receive question sentence text with semantics and overcome differences in sentence structures that have the same meaning, synonyms and word variations to provide relevant answers using the Sentence-BERT model. In this system, the knowledge base used is FAQ from several websites such as ipcc.ch, climate.nasa.gov, nature.org, imperial.ac.uk, natgeokids.com, theguardian.com. The results of experiments fine-tuning the indoSBERT model with 6461 data for training, 1384 validation data, and 1385 test data, as well as with hyperparameters, namely batch size 16, learning rate 2e-5, warmup steps 0.1, and with epoch trials 3, 5, and 10 respectively resulted in MRR evaluation values of 0.878, 0.882, and 0.880. Epoch 5 with the best evaluation results was selected for use in the system. The result of testing the system using 10 queries is that the model can understand the context of the question so that it can provide relevant answers at the top position for each question, with an MRR evaluation value of 0.9.57 PagesSkripsi Sarjan
