Prof.<\/span> Dursun Delen<\/span><\/a> is the Holder of William S. Spears Endowed Chair in Business Administration, Patterson Family Chair in Business Analytics, Dir<\/span>ector of Research for the Center for Health Systems Innovation, and also Regents Professor of Management Science and Information Systems in the Spears School of Business at Oklahoma State University (OSU). <\/span><\/p>\n Scott Fincher<\/span><\/a> is a data scientist on the Evangelism team at KNIME and one of the biggest contributors to the<\/span> KNIME Forum<\/span><\/a>.<\/span><\/p>\n Text Mining is currently experiencing a surge in popularity, mainly due to the development of more advanced chatbots, advances in deep learning architectures applied to free text generation, and the abundance of text data generated every day from web applications, e-commerce, and social media.<\/span><\/p>\n In this interview, I would like to dig deeper into some common problems that data scientists face when analyzing text documents. I will be talking to Dursun and Scott about blending data sources, language specific processing, minimum amounts of data, deployment, and other commonly posed questions.<\/span><\/p>\n [Rosaria]<\/i><\/b> I wo<\/span><\/i>uld like to start from the beginning of all data science projects. Where can I get data for a text mining project? In particular, how<\/span><\/i> would you proceed if the information is not contained in one location only, but distributed across many websites and blog posts?<\/span><\/i><\/p>\n [Scott]<\/b> There are many repositories for text data, often organized around specific topics. One great data repository for beginners is probably<\/span> Kaggle<\/span><\/a>. There are a number of datasets in there that can be used to take the first steps in text mining.<\/span><\/p>\n Now for the second question. Depending on where your da<\/span>ta is located, you have several options for bringing them into a KNIME workflow. <\/span><\/p>\n For flat files, you can use the<\/span> File Reader<\/span><\/a> node as a start for individual files, and the<\/span> Tika Parser<\/span><\/a> node for reading large groups of documents – for example if you have folders full of Word or PDF files. The Tika Parser node, especially, is very flexible and can read a large variety of data files and formats. <\/span><\/p>\n If your data is stored in a database, KNIME Analytics Platform has nodes to access most of them via the<\/span> Database Extension<\/span><\/a>. <\/span><\/p>\n You can also access data stored on web sites using the<\/span> Webpage Retriever<\/span><\/a>, or via<\/span> REST API<\/span><\/i> with a<\/span> GET Request<\/span><\/a> node. There are even nodes for accessing tweets directly from the<\/span> Twitter API<\/span><\/i>. One of the strengths of KNIME Analytics Platform is its ability to pull data in and blend it from a wide variety of sources.<\/span><\/p>\n<\/p>\n [Rosaria]<\/i><\/b> <\/span>Do you have examples for grabbing data, for example from pdf and docx files or from web scraping?<\/span><\/i><\/p>\n [Scott]<\/b> Sure! We have several workflows available to help you get started. As an example, we have a workflow that demonstrates accessing data from both local MS Word files and remote web pages, and how you might combine the two together. It’s available on the <\/span> KNIME Hub<\/span><\/a> as “<\/span>Will they blend? MS Words meets Web Crawling<\/span><\/a>” (<\/span>https:\/\/kni.me\/w\/igO2cb4VhPHvmiQF<\/span><\/a>), along with lots of other examples. The<\/span> KNIME Hub<\/span><\/a> is a place to access workflows, components, and extensions built by both KNIME and the wider user community. You can search on keywords for particular use cases, algorithms, or whatever is of interest to you. Maybe the best part is that, if you come up with a really great solution that you think would benefit everyone, you can upload it yourself and share it with the community!<\/span><\/p>\n<\/p>\n [Rosaria]<\/i><\/b> <\/span>And what about connecting to repositories on the cloud, like S3?<\/span><\/i><\/p>\n [Scott]<\/b> We also have several nodes available to help you easily access cloud resources, whether you prefer Amazon, Azure, or Google Cloud. Because accessing different sources is such a common need, we recently updated and published a collection of blog posts in the<\/span> “Will they blend?”<\/span><\/a> booklet on our<\/span> KNIME Press<\/span><\/a> site. This book is freely available, and focuses on interesting ways to combine all kinds of datasets from a variety of sources, with plenty of example workflows for you to download and explore.<\/span><\/p>\n [Rosaria]<\/i><\/b> <\/span>Let’s move on to the next step in a text mining project: text processing. What does “tokenization” refer to? <\/span><\/i><\/p>\n [Dursun]<\/b> Tokenization is the process of separating a textual document into its low-level units, such as words, numbers, characters, etc. <\/span><\/p>\n<\/p>\n [Rosaria]<\/i><\/b> <\/span>What tokenizer should I use then?<\/span><\/i><\/p>\n [Dursun]<\/b> The basic tokenizer algorithms are available in the<\/span> Strings to Document<\/span><\/a> node (OpenNLP English Word Tokenizer, OpenNLP WhiteSpace Tokenizer, Stanford NLP Tokenizer, etc.). Some of these simple tokenizers are English language specific, some are language agnostic white-space driven. More sophisticated tokenization algorithms can be found only for some languages. My recommendation is use the basic ones, if nothing else is available.<\/span><\/p>\n<\/p>\n [Rosaria]<\/i><\/b> Is there a node that can remove punctuation characters? <\/span><\/i><\/p>\n [Scott]<\/b> Yes – it’s the<\/span> Punctuation Erasure<\/span><\/a> node. This is one of the easiest nodes to use in the<\/span> KNIME Textprocessing<\/span><\/a> extension because there’s almost no configuration for it – you just apply it to your documents and continue with your processing. You may find sometimes that you still have hyphens, for example, in some of your tokenized words even after punctuation removal – if that’s the case, you may need to revisit your tokenizer algorithm selection.<\/span><\/p>\n<\/p>\n [Rosaria]<\/i><\/b> How do you create a dictionary, like for example a stop word dictionary?<\/span><\/i><\/p>\n [Scott]<\/b> The<\/span> Stop Word Filter<\/span><\/a> node comes with its own internal dictionary for several languages, but if you want to provide your own, that’s no problem. Just create a table consisting of a single column, with one stop word per row, and feed that into the optional input port of the node. You can use this same approach to create custom dictionaries for other nodes that might need them – for example, the<\/span> Dictionary Tagger<\/span><\/a> node.<\/span><\/p>\n [Rosaria]<\/i><\/b> This is a classic question: What are the most common use cases requiring text mining?<\/span><\/i><\/p>\n [Dursun]<\/b> Probably the most popular use cases for text mining nowadays are<\/span> Sentiment Analysis<\/b> and<\/span> Topic Detection<\/b>. Use of text mining to extract knowledge from<\/span> social media<\/b> (tweets) or from<\/span> customer reviews of products and services<\/b> are very popular. Also of interest is<\/span> classification of documents<\/b> – this is similar to the supervised version of sentiment analysis, but the classes are not just limited to degrees of positive and negative feeling. One area where text mining is showing up in academic circles is<\/span> Literature Mining<\/b>. Here, the idea is to discover latent topics from a collection of papers and then look at their dominance overtime and at their relationship to journals and disciplines. <\/span><\/p>\n<\/p>\n [Rosaria]<\/i><\/b> Would you recommend a few use cases for beginners?<\/span><\/i><\/p>\n [Scott]<\/b> Sure!<\/span> Sentiment analysis<\/b> is a good starting point. Conceptually it’s fairly easy to grasp, but at the same time can be treated with different degrees of complexity: running a<\/span> lexicon based analysis<\/span><\/a>, implementing a classic<\/span> supervised machine learning approach<\/span><\/a>, or for the more experts using a<\/span> deep learning network<\/span><\/a>. These three linked examples all analyze IMDB movie reviews and try to classify them as positive or negative.<\/span><\/p>\n Another use case that could be approached by a beginner is<\/span> topic extraction<\/b>, maybe with the help of the LDA algorithm. Here is an example from the KNIME Hub titled “<\/span>Topic Extraction<\/span><\/a>”, using the Parallel LDA node to mine topics from Tripadvisor restaurant reviews.<\/span><\/p>\n If you’re interested in importing data from Twitter and doing some exploratory visualization using a<\/span> word cloud<\/b>, this other simple workflow, titled “<\/span>Interactive Tag Cloud from Twitter Search<\/span><\/a>”, can show you how that might be done.<\/span><\/p>\n<\/p>\n [Rosaria]<\/i><\/b> Let’s take topic detection, then. What is the most common algorithm used to extract topics (unsupervised) from a text?<\/span><\/i><\/p>\n [Dursun]<\/b> This is probably LDA, where LDA stands for Latent Dirichlet Allocation. LDA produces a pre-determined number of latent (or hidden) topics from a collection of documents. The output is two-fold: first, a distribution of topics that defines a document, and second, a distribution of words\/terms that defines a topic. The<\/span> Topic Extractor (Parallel LDA)<\/span><\/a> node produces both of these outputs as separate tables.<\/span><\/p>\n<\/p>\n [Rosaria]<\/b> Talking about LDA, here’s a question that comes up often. How large does the document collection need to be to apply LDA? <\/span><\/i><\/p>\n [Dursun]<\/b> The short answer is “<\/span>large enough but not too large<\/span><\/i>”. If the computational power is not an issue, and the documents are coming from a representative application domain, then larger is better. As a rule of thumb, in text mining and in topic detection, large quantities of smaller size documents tend to produce more reliable and consistent results.<\/span><\/p>\n <\/span><\/p>\n [Rosaria]<\/b> Are there alternatives to LDA, in case LDA has limited success?<\/span><\/i><\/p>\n [Dursun]<\/b> LDA has been the latest and greatest of all topic detection methods until recently. Nowadays there are newer methods like Word2Vec, WordEmbedding and Deep Learning (using RNNs\/LSTMs) that take text mining and topic modeling to a new dimension, by including the contextual\/positional information from the sequential nature of language. About 15 years ago, I used simple clustering on raw term-document matrices (TDM) and then repeated it on SVD (Singular Value Decomposition) values for literature mining. It produced pretty good results for that time and the<\/span> study was published in a high impact academic journal<\/span><\/a>. <\/span><\/p>\n<\/p>\n<\/a>Meet two text mining experts in today’s interview, which explores some of the common issues faced by data scientists in text analytics. Prof. Dursun Delen<\/a> and Scott Fincher<\/a> are the teachers of the “<\/span>[L4-TP] Introduction to Text Processing<\/span><\/a>” course regularly run by<\/span> KNIME<\/span><\/a>. This course is based on the<\/span> KNIME Text Processing Extension<\/span><\/a>.<\/span><\/p>\n
<\/a>Collecti<\/span>ng the Data<\/span><\/h2>\n
Text Processing and Dictionaries<\/span><\/h2>\n
Use Cases<\/span><\/h2>\n