KDnuggets Home » News » 2014 » Jun » Opinions, Interviews » Interview: Samaneh Moghaddam, Applied Researcher, eBay on Opinion Mining – Typical Projects and Major Challenges ( 16:n16 )

Interview: Samaneh Moghaddam, Applied Researcher, eBay on Opinion Mining – Typical Projects and Major Challenges


We discuss typical sentiment analysis problems at eBay, underrated challenges, career motivation, important soft skills and more.



SamanehSamaneh Moghaddam is part of the “Customer Connect Data Science Engineering” team at eBay that transforms customer service data into actionable information. She is working on topic and sentiment models to identify and summarize customers’ pain points from user feedback. Samaneh Moghaddam holds a PhD in Computer Science with a thesis on aspect-based opinion mining. She has published several papers in the area of sentiment analysis in the top international conferences such as WWW, ACM SIGIR, ACM CIKM, and ACM WSDM.

First part of interview.

Here is second part of my interview with her:

Anmol Rajpurohit Q4. What are the kind of problems that you work on at eBay?

Samaneh Moghaddam: At eBay, I am part of the Customer Connect team. The Customer Connect team is in charge of transforming customer service data into actionable information. I am mainly working on sentiment analysis models to identify and summarize customers’ pain points from user feedback. The project that I am working on currently is mining and summarizing actionable information from eBay App reviews. Our goal is to make the experience easier and more favorable for our app customers.

AR: Q5. What are the most underrated challenges of working on Sentiment Analysis?

ChallengesSM: Sentiment Analysis or opinion mining represents a large problem space. Sentiment analysis can be done at document-level (e.g., opinion spam detection, opinion quality and helpfulness estimation, etc.), or at sentence-level (e.g., opinion question answering, opinion mining in comparative sentences, etc.) or at the phrase level (e.g., aspect-based opinion mining). Some challenges like dealing with noisy information or co-reference resolution are common in the whole problem space. However, each specific problem has its own challenges.

One of the main challenges in the area of aspect-based opinion mining is identifying implicit aspects and sentiments. There are usually many types of implicit aspect expressions in a review. Adjectives and adverbs are perhaps the most common types because most adjectives describe some specific attributes or properties of items, e.g., expensive describes ‘price,’ and beautiful describes ‘appearance.’ Implicit aspects can be verbs or complex phrases too. Although there have been some works considering extraction of implicit aspects, further research is still needed.

In the same way, while most sentiments are expressed through adjectives and adverbs, nouns (e.g., rubbish, junk, and crap) and verbs (e.g., hate and love) can also be used to express sentiments. Apart from individual words, there are also sentiment phrases and idioms, e.g., ‘cost someone an arm and a leg’. While identifying these types of sentiments is very difficult, the main challenge is predicting the polarity/rating of them.


AR: Q6. What motivated you to work in the field of Data Science?

MotivationSM: I got my PhD in computing science and I always enjoyed finding and solving challenging problems. In fact, finding and defining a problem satisfy my curiosity and attacking and solving it satisfy my creativity. I usually take time to learn what needs to be solved, and how it will be used and by whom. When I come up with a clear problem definition, I attack it by studying the relationships between the data. Finally, I attempt new problem-solving approaches and potential solutions.

AR: Q7. What is the best advice you have got in your career?

SM:
Settle for a job that keeps your passion alive! A place where you can apply your knowledge, use experience and at the same time learn new topics and skills.


AR: Q8. What skills do you think are the most important for practitioners in the field of Data Science?

Building SkillsSM: Strong background in statistics, algorithms and machine learning as well as knowledge of diverse technologies such as Hadoop, Java, Hive, Pig, Python, R, etc. Innovation, problem-solving skill and presentation ability are also must-have. Depending on the topic that one is working on, other skills and knowledge will be necessary. For example, a scientist in the area of sentiment analysis/opinion mining, needs to also have strong background in natural language processing, information retrieval and text mining.

AR: Q9. On a personal note, are there any good books that you’re reading lately and would like to recommend? What do you like to do when you are not working?

SM: I recently read “How to talk so kids will listen & listen so kids will talk”. I am a new parent and I found this book really helpful. I recommend every new parent to read this book.
I love spending time with my family. I also like reading short stories, biking and hiking.

Related: