Semantic text similarity (STS) uses specific test collections as its performance evaluation measurement. The test collections consist of text pairs with the same meaning even though in different text form. The existence is scarce compared with information retrieval (IR) test collections. This paper investigates the possibility to reuse IR test collections for STS tasks. Text pairs are derived from the relevant pair of IR test collections. Latent semantic analysis (LSA) and explicit semantic analysis (ESA) evaluate Glasgow's test collections, which are provided by ACM SIGIR community. Jaccard index measures the lexical similarity. Recall metric measures retrievability of recycling test collection with two existing test collections, Microsoft research paraphrase corpus and Microsoft research video description corpus, as evaluation baselines. Evaluation yields a promising outcome; the evaluated test collections have low Jaccard index and their recall values between the two baselines.

No abstract

<p class="Abstrak">Pada sistem temu kembali informasi berbentuk teks maupun <em>text mining</em>, terdapat proses pengindeksan. Teks diproses dengan tujuan mengintisarikan informasi berbentuk teks tersebut. Salah satu proses yang dilakukan adalah <em>stopword filtering</em>,<em> </em> beberapa kata yang tidak layak diindeks diabaikan berdasar sebuah daftar. Di dalam sistem berbahasa Indonesia, terdapat beberapa versi daftar <em>stopword</em> yang tersedia bebas. Penelitian ini bertujuan mengevaluasi daftar yang telah tersedia tersebut. Tujuan akhir dari penelitian ini adalah telaah daftar yang tersedia berdasarkan tata bahasa Indonesia, cara penyusunan, dan kebiasaan perambah internet. Dari hasil telaah diperoleh fakta bahwa daftar yang tersedia dibangun dengan analisis frekuensi kemunculan kata pada sebuah korpus (<em>corpus</em>) teks, tanpa memperhatikan jenis kata ataupun kebiasaan pengguna internet. Hasil lain penelitian ini  adalah beberapa rekomendasi lebih lanjut bagi para peneliti di bidang ini ketika membutuhkan daftar <em>stopword </em>bahasa Indonesia, yaitu daftar yang memperhatikan jenis kata dan kebiasaan pengguna internet melalui mesin perambah yang tersedia.</p><p class="Abstract"><em><strong>Abstract</strong></em></p><p class="Abstract"><em>Most of text-based information retrieval system uses indexing process. The system processes the texts in order to obtain the information essence. One of the process is stopword filtering, several words are being ignored based on a stopword list. Several Indonesian stopword list are available openly. Therefore, this paper evaluates the available lists based on Indonesian formal grammar, its preparation technique, and internet surfer habit. The results show all of the list are developed by term frequency analysis based on a text corpus. This paper also provides several recommendations for researcher both in text mining and text-based information retrieval field, developing stoplist by the word type and internet surfer habit.</em></p>

Website and blog are popular as a media to spread news. The validity of an article of news’s can either be valid or fake. A fake article of news is usually called a hoax news article. The purpose of making hoax news is to persuade, manipulate, affect to people to do something that contradicts or prevents the right action. A hoax news usually used threats or misleading information to make them believe things that are not real. This research proposes an experiment using naïve Bayes to detect hoax news in Bahasa Indonesia. In this research, we use our own dataset consisting of a total of 600 valid and hoax articles. We asked three reviewers to conduct manual classification for our dataset. Final tagging was obtained by adopting the maximum score from the three reviewers. In our experiment, we show that naïve Bayes can classify Indonesian online news articles with term frequency feature using the PHP-ML library component’s. We obtained an accuracy is 82.6% with static testing and 68.33% with dynamic testing. We give free access to the dataset so the future research can replicate, comparing the result and make a baseline testing.Keywords : Hoax News Detection, Naïve Bayes Classifier.