Sentiment Analysis of Islamic Waqf: Evidence in Indonesia

Transcription

1. SENTIMENT ANALYSIS OF ISLAMIC WAQF : EVIDENCE IN INDONESIA AAM SLAMET RUSYDIANA Shariah Economic Applied Research and Training Indonesia E-mail: aamsmart@gmail.com Abstract It is important to do research on public sentiment towards waqf presence in a country in order to know public response to its existence. This study aimed to determine public sentiment towards waqf in Indonesia. Data were collected from 80 articles, journals and other writings. Data were analyzed using the software Semantria as an analytical tool in the form of text. The results showed that the assessment of existence of waqf in Indonesia amounted to 66% of the community showed positive and high positive sentiment, 11% indicate negative sentiment and 23% indicates a neutral sentiment. Therefore, stakeholders need to take advantage of the awakening momentum of waqf in Indonesia so that in the future they can be a solution to the problems of social economy and the benefit of society. Keywords: Islamic Waqf, Sentiment Analysis, Social Finance INTRODUCTION Islamic finance in Indonesian emerged around 1992, spearheaded by Bank Muamalat Indonesia. Once it starts growing Islamic Banks, Islamic Business Unit, Syariah Rural Bank, Syariah Cooperation, Takaful, Islamic pawnshop, Wakaf, Zakat and Islamic Financial Institutions. Taken together, the Islamic financial system and conventional financial system synergistically support the mobilization of public funds more widely to improve funding for sectors of the national economy. The function and role of Islamic financial economy and the financial system is more than expected, especially the experience of the financial crisis has been reviving the essence of the activity of financial institutions (Nurfalah et. al., 2018). Economic and Islamic finance, has been based in concept to real transactions which does not ignore the practice of speculation and financial fundamentals and the real, so as not to cause any bubble in the economy and financial system. Economic and Islamic finance system are present based on the achievement of justice and the distribution of economic prosperity and ethical values that are universal, that is acceptable to all parties. In practical terms, in the end the economic and Islamic finance provides a clear line of connection that product and financial transactions that occur in the market should be highly relevant and directly related to the real sector (FSA, 2013). Apart from Islamic finance, no less important is the existence of Islamic social finance. An important instrument of Islamic social finance is waqf. Waqf is meant as the legal action of wakif (someone who takes waqf) to separate and or to give half of his/her property
2. 124 & Maqdis : Jurnal Kajian Ekonomi Islam - Volume 3, Nomor 2, Juli – Desember 2018 to be utilized forever or in some certain period which is an integral part of the analysis of based on his/her interests for prayer need or text mining. general interest based on sharia. Sentiment analysis or opinion mining is Based on Kahf and Mahamood (2011), the process of understanding, extract and waqf is divided info into some parts based process the textual data automatically to get on the aims which are: Wakaf Khairi or waqf the sentiment of information contained in an which is utilized for general society; Wakaf Ahli opinion sentence. Sentiment analysis is done or waqf which is utilized for family or relative to see opinions or opinions tendency towards a and descendents without differentiate between problem or an object by a person, whether or the rich and the poor, the health or the sick among opine tend to view negative or positive. One them; and Wakaf Musytarak or waqf which is example of the use of sentiment analysis in utilized for general society and family. the real world is the identification of market Waqf is one of Islamic laws related to trends and market opinion to an object goods. people’s lives as ijtima’iyyah worship for public The magnitude of the effect and benefits of interest as a devotion to Allah SWT (Fathurrohman, sentiment analysis led to research and sentiment 2012). Moreover, Islam has institutional preparation analysis based applications is growing rapidly. to acquire fund so the poor can be independent. Even in America, there are about 20-30 companies For this purpose, waqf can be done in addition that focus on sentiment analysis service. to alms mandatory payment and other voluntary Sentiment analysis or opinion mining is the payment contributions. So if waqf is developed computational study of the opinions of people, and managed productively, it can be an alternative appraisal and emotions through entities, events for poverty reduction (Rusydiana & Devi, 2017). and attributes of (Biu, 2010). The basic task in Given the very large role of Islamic social sentiment analysis is classifying the polarity finance especially waqf in supporting the of the text in the document, sentence, or features Indonesia economy in the future, it is necessary /aspects level - whether the opinions expressed to review the public response to waqf and cash in the document, a sentence or entity features/ waqf issue in Indonesia over the years. aspects are positive, negative or neutral (Dehaff, Therefore, as a material assessment of the 2010) importance of waqf, it would require an In wikipedia, text mining, refers to the assessment of public sentiment against waqf process of taking high-quality information from sentiment and sentiment analysis can be used text. High quality information is usually obtained
3. Sentiment Analysis of Islamic Waqf (Aam Slamet Rusydiana) & 125 through forecasting patterns and trends through this case, General Directory only receives 5% means such as learning statistical patterns. Text out of net income of waqf as the supervising mining typically involves structuring processes and auditing costs. General Directory is text input (usually a parsing, along with the appointed by Prime Minister and it is under addition of some linguistic features derivative the supervision of Prime Minister office. The and removal of some of them, and subsequent services given by the General Directory of Waqf insertion into the database), determine patterns include health, education and social services. in structured data, and finally evaluate and Moreover General Directory of Waqf is also interpret the output. 'High quality' in the field of doing some relationship and investment in text mining usually refers to some combination some other institutions such as: Ayvalik and of relevance, novelty, and interestingness. Aydem Olive Oil Corporation, Tasdeleln Healthy Typical text mining process include text Water Corporation, Auqaf Guraba Hospital, categorization, text clustering, extraction concept/ Taksim Hotel (Sheraton), Turkish Is Bank, Aydin entity, the production of granular taxonomies, Textile Industry, Black Sea Copper Industry, sentiment analysis, inference documents, and Construction and Export/Import Corporation, entity relationship modeling (ie, learning the Turkish Auqaf Bank. Turkish Awqaf Bank is relationship between the entities named). established by General Directory in 1954. General LITERATURE REVIEW Directory has some stocks in the bank about Contemporary Waqf 75%. This bank is one of the biggest bank in On the contemporary era, waqf in some Turkey with the total capital of 17 Billion TL countries have been developed in many public (45 million USD), this bank has 300 branches sectors with multi-benefit. Turkey as the example, all over Turkey. The profit received in 1983 is 2 based on Hasanah (1997), waqf in Turkey is billion TL (5 million USD). The income from managed by Waqf General Directory and there this bank is utilized as the management, the are also managed by mutawalli. Besides managing fixation, and other need of property waqf. waqf, Waqf General Directory has also done Unlike in Turkey, waqf in Egypt initially some supervising and controlling through the was mostly included in expert waqf (waqf for managed waqf by mutawalli or new waqf (Art 78 family) and khairi waqf (waqf for public). In Civil Law). In the regulation laws in Turkey, expert waqf, wakif can take his or her property waqf has to have management council. Waqf in that he/she has used as a waqf or changing Turkey has to be audited every two years. In the target, but he or she cannot take for his/her
4. 126 & Maqdis : Jurnal Kajian Ekonomi Islam - Volume 3, Nomor 2, Juli – Desember 2018 interests. In khairi waqf, wakif cannot take his or Sentiment Analysis her property and cannot change the target. It Sentiment analysis or opinion mining refers is because many problems appear in the expert to a broad field of natural language processing, waqf implementation, at the end the expert computational linguistics and text mining. waqf is erased and automatically also erase In general, it aims to determine the attitude muaqqat waqf (waqf which is limited by time) of the speaker or writer with respect to a because previously in Egypt, muaqqat waqf was particular topic. Attitude may assessment or only in expert waqf. evaluation of them, a statement of their affective In 1946, Egypt government released Laws (emotional statement authors when writing) number 48 in 1946 about Waqf Laws. These or the intended emotional communication laws have contents about waqf laws such as (emotional effect the author wants to readers). when is waqf happened, waqf requirements, The basic task in sentiment analysis is classifying people who own the waqf property, waqf nazhir, the polarity of the text in the document, sentence, the power of nazhir towards property and its or features/aspects level - whether the opinions development. On that laws in verse number expressed in the document, a sentence or entity 8 is also stated whether the non-statical waqf features/aspects are positive, negative or is allowed or not, company stocks that the neutral (Dehaff, 2010), Further sentiment analysis company is allowed based on the Islamic Sharia. can be expressed emotionally sad, happy, or In Egypt, the waqf issues are obligated angry. in detail and it is always developed based on the Some studies classify the polarity of the existing development. As the example in document on a scale of multi-directional, Egyptian waqf Laws, it is also managed about attempted by Pang & Lee (2005) and Snyder farm land changing which is being waqf for & Barzilay (2007) include: extending the basic the good aim and the Economical Foundation is task of classification review of the movie as given the authority for developing land farm a positive or negative to predict a good star waqf result to develop economical humanity. rating scale 3 or 4, while (Snyder & Barzilay, The regulation about this waqf continuously 2007) conducted in-depth analysis on a restaurant is revised with situation and condition also review, predicting the ratings for various aspects always based on the Islamic Sharia, so in 1971 of the restaurant are given, such as the food the institution is made to handle the waqf and atmosphere (in a five-star scale). and its development. A different method to determine the sentiment
5. Sentiment Analysis of Islamic Waqf (Aam Slamet Rusydiana) & 127 is the use of large-scale systems where the words We can track their products, brands and commonly associated sentiment negative, neutral people, for example, and determine whether or positive with those given a number on a scale they are positive or negative views on the of -5 to +5 (most negative to the most positive) web. It allows a business to keep track of: and when a piece of structured text analyzed Perception of a new product, brand perception, natural language processing, concept further reputation management and other related analyzed to understand these words and how issues. Expression or sentiment refers to the they relate to the concept. Each concept was focus of specific topics, a statement on the then given a score based on how words relate to topic may be different meanings to the same the concept of sentiment, and the scores were statement on different subject. For example, related. This allows the movement to a more is a good thing to say the flow of the film are sophisticated understanding of sentiment based not predictable, but it is not a good thing if on a 11 point scale. the 'unpredictable' is stated on the steering wheel Research in a different direction is the of the vehicle. Even on a particular product, identification of subjectivity/objectivity. This the same words can describe the meaning of the task is usually defined as classify a given text opposite, examples are a bad thing for a start-up (usually a sentence) into one of two classes: time on the digital camera if it is declared "old", objectively or subjectively (Pang & Lee, 2008). but if the "old" age stated on the battery it will be This problem can sometimes be more difficult a positive thing. Therefore, in some studies, than the classification polarity (Mihalcea, et., especially on the product review, work was al, 2007) subjectivity words and phrases may preceded by defining the elements of a product depend on the context and objective document that is being discussed before starting the process may contain subjective words (for example, of opinion mining (Barber, 2010). a news article quoting the opinion of people). First thing in the processing of documents is In addition, as mentioned by (Su & Markert, to break down a group of characters in the 2008), the results are highly dependent on the word or token, often referred to as tokenisasi. definition of subjectivity used when annotate Tokenisasi is complex for computer programs as text. However, (Pang & Lee, 2004) shows that some characters can be found as token delimiters. removing an objective sentence of a document Delimiter is the character of spaces, tabs and before classifying polarity help improve new line "newline", while the character () <> performance. !? Sometimes used as a delimiter but sometimes
6. 128 & Maqdis : Jurnal Kajian Ekonomi Islam - Volume 3, Nomor 2, Juli – Desember 2018 not depends on the environment (Wulandini Determinants Of People On Public Opinion. & Nugroho, 2009). Mentioned that the increasingly widespread Empirical Studies use of social networks like Twitter makes social Research on text mining or sentiment networking such as very large data. This study analysis has been carried out. As for some was to determine the opinion or sentiment of summaries of previous studies related to the social networking users of a topic. One important implementation of opinion mining (text mining) topic is a public figure, as a candidate for as follows: governor, party chairman. The results of the Rozi, et., al. (2013) examined the Public Opinion analysis and testing shows the preprocessing Data Extraction In Higher Education. In this stage does not have a significant effect on the research, opinion mining system developed accuracy (69.4% - 72.8%) sentiment classification. to analyze public opinion in college. In the As for the extraction topic show that use of document subprocess subjectivity and detection Tf-Idf with cumulative discounted able to targets used Part-of-Speech (POS) Tagging increase the amount of extracted corresponding using Hidden Makov Model (HMM). On the topics. However, it has a weakness when facing results of the process of POS Tagging then topics contained in almost the entire interval applied the rule to determine whether a or a topic that is not sourced from news in the document including opinions or not, and to internet media. know which part of the sentence which is the Faishol (2011) examined the implementation object of the target opinion. Documents that of Text Mining to Support Search Topics In are recognized as opinion is further classified the e-library Using Mobile Device. This study into positive and negative opinion (Opinion uses text mining methods that implement the subprocesses orientation) using a Naive Bayes algorithm cosine similarity to peringkatan document classifier (NBC). From the test values obtained (page rank). This is necessary because of the precission and recall for subprocesses document amount so large collection of documents owned subjectivity is 0.99 and 0.88, for a target by a library, we need a method to peringkatan detection subprocesses are 0.92 and 0.93, and for these documents when requested. In text mining the opinion subprocesses orientation are 0.95 there are several important processes, namely and 0.94. the folding case, tokenizing, filtering, and Sunni and Widyantoro (2012) analysis analysis stemming. Stemming used is Porter's on sentiment and sentiment extraction Topics algorithm Indonesian-language text and the
7. Sentiment Analysis of Islamic Waqf (Aam Slamet Rusydiana) & 129 analysis in the document weighting algorithm, show that the SVM method provides better TF/IDF and cosine similarity (Vector Space Model). performance than either method on NBC for Test data obtained from the central library of classifying opinion positive opinion in English the State Islamic University of Malang in the and Indonesian language. While NBC gives form of abstraction thesis. From the test results better performance in classifying test data of obtained that relevant documents are received negative opinion in Indonesian language. by users reaches 100% and accuracy of data Rusydiana et. al. (2018) analysis on sentiment relevant to the received data users reached of microtakaful industry in Indonesia and an average of 78.2%. Malaysia. Based on the results of text mining Saraswati (2011) examined the Text Mining analysis of writings, articles and journals Methods Naïve Bayes classifier and Support about microtakaful in Indonesia and in Malaysia Vector Machines for Sentiment Analysis. Text it can be concluded that: The majority of the mining is concluded that, referring to the positive sentiment in Indonesia against micro process of taking high-quality information from takaful attendance was 52% while in Malaysia as text. High quality information is usually much as 62%. So that the positive sentiment obtained through forecasting patterns and trends in Malaysia is greater than in Indonesia. The through means such as learning statistical negative sentiment on microtakaful presence patterns. Typical text mining process include in Indonesia as much as 28%, while in Malaysia text categorization, text clustering, extraction as much as 23%. So that the negative sentiment concept/entity, the production of granular in Malaysia less than in Indonesia. Sentiment taxonomies, sentiment analysis, inference Neutral on microtakaful presence in Malaysia as documents, and entity relationship modeling. In much as 15%, while in Indonesia by 20%. this study discussed the classification of opinions The research about waqf has done by as a positive opinion and a negative opinion Rusydiana and Devi (2018). This research is on the data and the data speak English Indonesia aim to identify the priority factors that being using Naïve Bayes classifier (NBC) and Support barrier to develop the practice of cash waqf in Vector Machine (SVM). Neither NBC nor the Indonesia using Analytic Network Process (ANP) method of SVM method provides good method. Result show that the problems appeared performance in sentiment analysis, opinion in managing cash waqf in Indonesia divided classification in English and Indonesian into 4 important aspects, there are: Human language in this study. The experimental results Resource aspect, trust aspect, system aspect, and
8. 130 & Maqdis : Jurnal Kajian Ekonomi Islam - Volume 3, Nomor 2, Juli – Desember 2018 sharia aspect. The rank for most priority problems derived from the reference or previous research. to less priority based on the priority result are: 1) Secondary data used in this study consisted of trust problems (whereas the most priority for 80 specific documents, either in the form of articles this sub-criteria is donators’ lack of trust), 2) sharia problems (is unfulfilled waqf covenants), 3) human resource problems (is misappropriation of waqf funds, 4) system problems (is weak of management systems). Strategies that can and journals related waqf in Indonesia. To support the strengthening of the analysis, then added the opinion of an expert waqf in Indonesia. It is intended to determine expert opinion on the results of the analysis in this study. The methodology used in this study is a sentiment be built to develop the practice of cash waqf in analysis or opinion mining. Sentiment Analysis Indonesia based on the priorities are: 1) more is a commonly used research to gauge public computerized cash waqf management, 2) the sentiment development of waqf education institutions, Analysis is a research branch in the domain 3) more comprehensive fund manager quality Text Mining boom that began in the early 2002's. improvement, 4) transparency and accountability His research began to flourish since the paper of in every step. Other research on waqf has done by Rusydiana and Al Parisi (2016). The result show that thestudies on waqf was still dominated by the discussion of the non-cash waqf rather than the cash waqf from 2011 to 2015, this provides an overview in general for researchers or experts to discuss more that related to the cash waqf. It has a great power in collecting funds from the community then it is used as a waqf productive for empowering the local economy. In addition, comparison of quantitative research method is still far less than the qualitative approach. This is the potential for increasing the waqf research using quantitative methods. RESEARCH METHOD In this study used secondary data, ie data on a theme issue. Sentiment Pang and Lee appear. Put simply, text mining to be intended for word processing and not process the numbers. Sentiment analysis is composed of three major subprocesses namely: Subjectivity Classification, Orientation Detection and Opinion Holder & Target Detection. Until now, most of the research in the field of sentiment analysis is aimed at English because it Tools/Resources for the English very much. Some resources are often used for sentiment analysis is SentiWordNet and WordNet. The basic task in sentiment analysis is classifying the polarity of the text in the document, sentence, or features/aspects level whether the opinions expressed in the document, a sentence or entity features/aspects are positive, negative or neutral (Dehaff, 2010), Further
9. Sentiment Analysis of Islamic Waqf (Aam Slamet Rusydiana) & 131 sentiment analysis can be expressed emotionally sad, happy, or angry. RESULT AND DISCUSSIONS that suit the needs of society. While the results showed 11% negative sentiment and the balance of 23% indicates a The author tried to calculate sentiment of neutral sentiment. This condition must be very Islamic waqf in Indonesia. As already known, reasonable considering the presence of waqf Sentiment Analysis is a commonly used research in Indonesia has find many challenges including to gauge public sentiment on a theme issue. As a Human Resource aspect, trust aspect, system, source of data, selected 80 specific documents, and sharia aspect. Based on Rusydiana & Devi either in the form of articles and journals related (2018) the most important problem in the to Islamic waqf in Indonesia. Tools used in development of waqf and cash waqf in Indonesia this research is Semantria as processing aids. is trust. This suggests that the waqf donators Results sentiment analysis of the condition of do not completely trust the waqf managers Islamic waqf in Indonesia can be seen in the to manage their donations in the form of cash. following figure: Some of the causes of the low trust in waqf managers could be the low quality of the waqf managers’ work performance, the occurrence of corruption and misappropriation of waqf funds, the lack of cash waqf education for donators and the divided sharia opinion of whether the cash waqf covenant is halal or not, etc. Figure 1. Waqf Sentiment Analysis in Indonesia From the picture above we can see that the majority view that the presence of waqf in Indonesia is very good (positive) precisely as much as 64% and 2% high positive. This means that the majority of the literature indicates positive sentiment. Therefore, this situation must be addressed properly by operators including associates. The social demand for Islamic waqf should be facilitated with products As for the solution priorities which are believed in developing waqf and cash waqf in Indonesia are: 1) sharia solutions (changing the term “tabarru’” to “grant funds”, 2) system solutions (waqf regulation/law support), 3) human resource solutions (conducting training programs about waqf), and 4) trust solution (dissemination). The strategies that could be employed in developing cash waqf are: 1) computerization of waqf fund management, 2) development of waqf education institutions, 3)
10. 132 & Maqdis : Jurnal Kajian Ekonomi Islam - Volume 3, Nomor 2, Juli – Desember 2018 improvement in waqf fund management quality, there are so many other sectors need to be and 4) transparency and accountability. development such as public sector. Resources Waqf issues recently has been governed and waqf fund sources will be much better if in Laws number 41 in 2004 about waqf which is it can support the government program in signed by the President in October 27th 2004, providing better public need. and it was included in the passage of Republic of If we talk about property waqf, the value of Indonesia in 2004 with number of 159 and land and property waqf in Indonesia alone also the additional passage of Republic of is estimated at around 4,4-billion-meter square Indonesia with number of 4459. This laws with the economic value of Rp.370 trillion (US$27 besides completing the recent waqf laws also billion). However, to date, most of the waqf regulate new issues as regulating waqf property land is limited to use in schools, mosques or has to be productived and the target has to be in public graves. Although the property of Waqf detailed such as helping the indigent, creating has benefited the community particularly in Indonesian waqf foundation, regulating money the surrounding areas. waqf, and other issues which are needed based on the contemporary development. Nowadays, the development of Waqf in Indonesia has been concerned by many scholars These law products has given the definite and Muslim philosophers, which triggers the law benchmark, public trust, and also protection born of new progressive ideas to develop Waqf. towards waqf asset. This legalization of laws This also encourages government to release is the strategical step to improve the role of regulation and guidelines to increase the waqf not only as the religion regulation, but empowerment program of Waqf. As the regulator, also as the potentially economical power to motivator, and facilitator, government embodies develop public wealthiness. Moreover with strategies step by step, from reinforcement of this legalized laws, the waqf objects will have Waqf institution, until tools and infrastructure wider scope not only statical properties but also supplying. non-statical properties such as money, precious CONCLUSIONS metal, certificates, rented rights, etc. Based on the results of text mining analysis It should be admitted that the development of writings, articles and journals about waqf of waqf in Indonesia in the past were focusing in Indonesia it can be concluded that the more to the non-productive waqf such as assessment of existence of waqf in Indonesia establishment of mosque and school. However, amounted to 66% of the community showed
11. Sentiment Analysis of Islamic Waqf (Aam Slamet Rusydiana) & 133 positive and high positive sentiment, 11% indicate Jiawei, H., Kamber, M., & Pei, J. (2012). Data negative sentiment and 23% indicates a neutral Mining: Concepts and Techniques. Massachusetts: sentiment. Therefore, stakeholders need to take Morgan Kaufmann. advantage of the awakening momentum of waqf Kahf, M, Mahamood, S.M. (2011). Essential in Indonesia so that in the future they can be Readings in Contemporary Waqf Issues. Kuala a solution to the problems of social economy Lumpur : CERT Publications. and the benefit of society. Suggestions presented Mihalcea, R. Banea, C. & Wiebe, J. (2007). in this paper are: It takes a number of articles to Learning Multilingual Subjective Language find out more on the existence waqf public via Cross-Lingual Projections. Proceedings of sentiment in Indonesia in order to become the Association for Computational Linguistics better analytical results. Research can be done (ACL), 976–983. also in other countries such as ASEAN, ASIA Miner, G., Delen, D., Elder, J., Fast, A., Hill, that in the World. T., & Nisbet, R. (2012). Practical Text REFERENCES Mining and Statistical Analysis for Non- Barber, I. (2010). Bayesian Reasoning and Machine Structured Text Data Applications. Oxford: Learning. USA: Cambridge University Press. Elsevier. Dehaff, M. (2010). Sentiment Analysis, Hard But Mumtaz, H., Asghar, S., & Rima, T. (2015). Worth It!. USA: Cambridge University Press. An Overview of Islamic Finance. International Faishol, M.A. (2011). Implementasi Text Mining Monetary Fund Working Paper June, 15(120). Untuk Mendukung Pencarian Topik Pada e- Nurfalah, I., Rusydiana, A.S., Laila, N., & Library Menggunakan Mobile Device. Malang: Cahyono, E.F. (2018). Early Warning to University of Islam Negeri Maulana Malik Banking Crises in the Dual Financial System Ibrahim. in Indonesia: The Markov Switching Fathurrohman, T. (2012). Wakaf dan Penanggulangan Kemiskinan (Studi Kasus Pengelolaan Wakaf di Approach. JKAU: Islamic Economics, 31(2): 133-156. Kabupaten Bandung Jawa Barat). Dissertation. Pang, B. & Lee, L. (2004). A Sentimental Unpublished. Jakarta: University of Indonesia. Education: Sentiment Analysis Using Hasanah, U. (1997). Peranan Wakaf dalam Subjectivity Summarization Based on Mewujudkan Kesejahteraan Sosial (Studi Kasus Minimum Cuts. Proceedings of the Association Pengelolaan Wakaf di Jakarta Selatan). Dissertation. for Computational Linguistics (ACL), 271– Unpublished. Jakarta: IAIN Jakarta. 278.
12. 134 & Maqdis : Jurnal Kajian Ekonomi Islam - Volume 3, Nomor 2, Juli – Desember 2018 Pang, B. & Lee, L. 2005. Seeing Stars: Exploiting Rusydiana, A. S., & Al Parisi, S. (2016). How Class Relationships for Sentiment Categorization Far Has Our Wakaf Been Researched?. with Respect to Rating Scales. Proceedings of Etikonomi, 15(1): 31-42. the Association for Computational Linguistics (ACL), 115–124. Pang, B., & Lee, L. (2008). Opinion Mining and Sentiment Analysis. Foundations and Saraswati, N.W.S. (2011). Text Mining Dengan Metode Naïve Bayes Classifier dan Support Vector Machines Untuk Sentiment Analysis. Denpasar: Program Pasca Sarjana Universitas Udayana. Trends® in Information Retrieval, 2(2): 1-135. Snyder, B. & Barzilay, R. (2007). Multiple Rozi, I. F., Pramono, S. H., & Dahlan, E. A. Aspect Ranking using the Good Grief (2013). Implementasi Opinion Mining Algorithm. Proceedings of the Joint Human (Analisis Sentimen) untuk Ekstraksi Data Language Technology/North American Chapter Opini Publik pada Perguruan Tinggi. of the ACL Conference (HLT-NAACL), 300–307. Jurnal EECCIS, 6(1): 37-43. Su, F., & Markert, K. (2008, August). From Rusydiana, A. S., Firmansyah, I., & Marlina, Words to Senses: A Case Study of Subjectivity L. (2018). Sentiment Analysis of Micro Recognition. Proceedings of the 22nd International Takaful Industry: Comparison Between Conference on Computational Linguistics, 1: Indonesia and Malaysia. International Journal 825-832. of Nusantara Islam, 6(1). Sunni, I, & Widyantoro, D.H. (2012). Analisis Rusydiana, A. S., & Devi, A. (2018). Elaborating Sentimen dan Ekstraksi Topik Penentu Cash Waqf Development in Indonesia Sentimen pada Opini terhadap Tokoh Using Analytic Network Process. International Publik. Jurnal Sarjana Institut Teknologi Journal of Islamic Business and Economics, Bandung Bidang Teknik Elektro dan Informatika, 2(1): 1-13. 1(2). Rusydiana, A. S., & Devi, A. (2017). Analisis Wulandini, F. & Nugroho, A. N. (2009). Text Pengelolaan Dana Wakaf Uang di Indonesia: Classification Using Support Vector Machine Pendekatan Metode Analytic Network for Web Mining Based Spat Ion Temporal Process (ANP). Al-Awqaf: Jurnal Wakaf dan Analysis of the Spread of Tropical Diseases. Ekonomi Islam, 10(2): 115-133. International Conference on Rural Information and Communication Technology, 189-192. View publication stats