Apache Lucene is an open source and free text search engine library written in Java. It is a technology suitable for applications that requires full-text search, and is available cross-platform.
N/A
IBM Watson Explorer
Score 8.4 out of 10
N/A
IBM Watson Explorer supports enterprise search with unstructured data analysis, machine learning, and content analysis to improve decision-making, support customer service or serve other business needs.
Apache Lucene offers great full-text search library that makes it easy to add search functionality to a website or other applications. Lucene is ideal if you want low-level access to the indexes and its APIs. For general purposes, Apache Solr, the web application built atop of Lucene can be used instead. Apache Solr comes with caching, HTTP/ JSON APIs and a simple web administration console.
The Watson Explorer is great because it potentially replaces a meriad of other low-level analytics products that we would need to use for data analytics and data mining. WEX isn't really suitable much beyond doing text and data analytics and performing machine learning, so if your team doesn't really have a use-case that fits all of these categories, it is worth looking at an alternative.
Free to try - It's possible to use most of the useful features of Watson Explore on their trial/demo accounts.
Super well-designed data analytics tool - Most of the tools and features of the explorer are really useful, and truly help you fully understand the depth of any format of textual data.
Extensive sources compatibility - WEX can retrieve data from a large range of sources, and the compatibility there is quite good as well.
We had difficulty porting the project to a cluster based environment on the cloud.
For our particular use case of retrieving documents based on text pattern matching, the program worked efficiently however, we did not find many resources for image pattern recognition based on their metadata.
Support is just OK, like most of the other IBM Watson products. The setup/integration is really hands-on, but it's also problematic because support later may take a considerable amount of time.
UI could still use a little more improvement - part of the administration and sources dashboards are hard to navigate.
The Application Builder is a great part of the product, but hard to learn/understand - this is where we needed the most support from IBM and tutorials/documentation.
The search and index performance of [Apache] Lucene is excellent and the quality of results is good, if not better. For implementing it with small scale applications it is a no brainer, Lucene is the best and most cost effective solution. Learning curve is not too steep either.
Google Cloud offers a Natural Language product, but it is just an API. This API doesn't offer the useful visualizations of relations, analytics, and graphs that IBM Watson Explorer offers on their interface. For this reason, we chose to go with IBM WEX. For later stages of our production, we decided to use Google's NLP API because we found that it was quick to integrate into production after studying data and developing models using IBM WEX.
Positive - Trial/demo period. This was really useful for us to figure out what features of WEX we liked most and how difficult it would be to integrate WEX into our workflow.
Negative - On-boarding was long and almost always requires support from IBM support, unlike most other products this advanced.
Positive - WEX replaced a large selection of alternative products we would have to use for the same functionality, and having all of that function in one place was definitely helpful.
