We performed a comparison between IBM SPSS Modeler, IBM Watson Explorer, and KNIME based on real PeerSpot user reviews.
Find out what your peers are saying about Knime, Weka, IBM and others in Data Mining."The most valuable features of the IBM SPSS Modeler are visual programming, you don't have to write any code, and it is easy to use. 90 to 95 percent of the use cases, you don't have to fine-tune anything. If you want to do something deeper, for example, create a better neural network, then you have to go into the features and try to fine-tune them. However, the default selection which is made by the tool, it's very practical and works well."
"It makes pretty good use of memory. There are algorithms take a long time to run in R, and somehow they run more efficiently in Modeler."
"The supervised models are valuable. It is also very organized and easy to use."
"It handles large data better than the previous system that we were using, which was basically Excel and Access. We serve upwards of 300,000 parts over a 150 regions and we need to crunch a lot of numbers."
"We have a local representative who specializes in SPSS. He will help us do the PoC."
"A lot of jobs that are stuck in Excel due to the huge numbers of rows are tackled pretty quickly."
"The quality is very good."
"In the solution, I like the virtualization of data flow since it shows what goes where, which is mostly the strength of the tool."
"Ease of use is pretty good as is the standardization of not actually having to have my own natural learning algorithms, just to use the Watson APIs."
"We take natural language that was happening in our repositories and our application and then feed it to the Watson APIs. We receive JSON payloads as an API response to get cognitive feedback from the repository data."
"For me, as a user, the most valuable feature is the ability to ingest and then retrieve information from a range of separate sources; the ability to dissect questions in context and actually answer them."
"I have found the auto-generated document very useful as well as the main keywords that are highlighted, which are used for the search functionality within IBM Watson Explorer."
"The ability to easily pull together lots of different pieces of information and drill down in a smarter way than has been possible with other analytics tools is key. Watson is all based on a set of AI and deep learning, machine-learning capabilities, and it is looking behind the scenes at some relationships that you likely would not have spotted on your own. It's pulling things together, categorizing some things, that are not something that you might have seen on your own."
"The valuable feature of Watson Explorer for us is data entities, and to see the hidden insights from within unstructured data."
"We are able to automate several functions which were done manually. I can integrate several data sets quickly and easily, to support analytics."
"This open-source product can compete with category leaders in ELT software."
"It has allowed us to easily implement advanced analytics into various processes."
"KNIME is quite scalable, which is one of the most important features that we found."
"Easy to connect with every database: We use queries from SQL, Redshift, Oracle."
"The solution is good for teaching, since there is no need to code."
"It is very fast to develop solutions."
"It also has very good fundamental machine learning. It has decision trees, linear regression, and neural nets. It has a lot of text mining facilities as well. It's fairly fully-featured."
"The product does not have a search function for tags."
"Requires more development."
"Unstructured data is not appropriate for SPSS Modeler."
"It would be beneficial if the tool would include more well-known machine learning algorithms."
"It is very good, but slow. The slowness may be because we have not finalized all the background information in SPSS. It still needs some tweaking."
"Time Series or forecasting needs to be easier. It is a very important feature, and it should be made easier and more automated to use. For instance, for logistic regression, binary or multinomial is used automatically based on the type of the target variable. I wish they can make Time Series easier to use in a similar way."
"I would like see more programming languages added, like MATLAB. That would be better."
"I understand that it takes some time to incorporate some of the new algorithms that have come out in the last few months, in the literature. For example, there is an algorithm based on how ants search for food. And there are some algorithms that have now been developed to complement rules. So that's one of the things that we need to have incorporated into it."
"Stability is actually one of the areas that could use improvement. Setting it up is always tough. Setting Explorer requires experts, but also the underlying platform is not that stable. So it really needs a good expert to keep it running."
"The solution is expensive."
"I would say, give some kind of a community edition, a free edition. A lot of companies do, even Amazon gives you some kind of trial and error opportunities. If they could provide something like that, it would be good."
"Small businesses will probably have a little harder time getting into it, just because of the amount of resources that they have available, both financial and time, but it really is a solution that should work for them."
"It is a little bit tricky to get used to the workflow of knowing how to train Watson, what can be provided, what can't be, how to provide it, how to import, export, and what it means every time you have to add a new dictionary"
"It needs better language support, to include some other languages. Also, they should improve the user interface."
"Much of IBM operates this way, where they have sets of tools that are in the middleware space, and it becomes the customer's responsibility or the business partner's responsibility to develop full solutions that take advantage of that middleware. I think IBM's finding itself in that spot with Watson-related technologies as well, where the capabilities to do really interesting and useful things for customers is there, but somebody still has to build it. Is that going to be the customer? Are they going to be willing to take on that responsibility themselves"
"More cognitive feedback would be good. The natural language analysis is great, the sentiment analyzers are great. But I would just like to see more... innovation done with the Watson platform."
"Compared to the other data tools on the market, the user interface can be improved."
"I would like it to have data visualitation capabilities. Today I'm still creating my own data visualtions tools to present my reports."
"The visualization functionalities are not good (cannot be compared to, for instance, the possibilities in R)."
"KNIME's documentation is not strong."
"The documentation needs a proper rework. "
"It could be easier to use."
"KNIME needs to provide more documentation and training materials, including webinars or online seminars."
"It's pretty straightforward to understand. So, if you understand what the pipeline is, you can use the drag-and-drop functionality without much training. Doing the same thing in Python requires so much more training. That's why I use KNIME."
Earn 20 points