Natural Language Processing, Sentiment Analysis, and Text Mining - Assignment Example

?Part One Sentimental Analysis Through sentiment analysis in tweeter, companies are able to acquire an effective and fast method to employ when monitoring the feeling of the entire public to their services or products. Through their posting in tweet as a heading of tweeter, it then study followers to learn the company`s consumers’ behaviors on extracting some information posted about them, their competitors and market. Additionally, evolution of tweeter API has enable sharing of material between website and tweeter without credential, but only a bit of interlinking coding. (Seif, Hassan, He, Yulani, Alani & Harith (2012). Consequently, sentimental analysis can be described as a tool for opinion mining. In this area, computations are involved and they end up in a natural language capable of processing opinions. It therefore give as detailed study into mood or emotion recognition, relevance computation, ranking, identifying source, giving genre distinction, and a summarization of the opinion (Pang & lee, 2008). Moreover, Pang & Lee (2008) states that in sentimental analysis texts need to be mapped to respective labels from a defined data set or through placing it from one end to another on continuum. Knowing above, a topic was to be investigated to determine the effectiveness of these analyses. The topic chosen included “saber tooth desktop” and in the rating, it occurred that 87% of the comment were positive and only 13% accounted for the negative. I downloaded it using sentiment140.com. However, reading through the tweets, it occurred to me that the facility was never accurate. It could never detect sarcastic messages, shortened words and it only analyzed the English part of the messages. Moreover, from the data of about four sets, correlation was sought out in order to determine the accuracy of the method and it was noted that for the positive tweets there was an F harmonic of about 6.5% and for the negatives it showed 4.2%. Consequently, this is an indication that sentiment gave a dependable result though they are not 100% accurate. Moreover, the above topic could be downloaded directly using tweeter API. One needs to have a tweeter account, create an application, from the application, one is offered machine readable consumer key access token, and consumer secrets and from those, one can receive tweeter updates on any website they specified and coded with the relevant details given when sign in for a tweeter API account. From the gotten results using tweeter Prolog-WordNet libraries, it indicated that there is a correlation between the sent tweets and the message they were conveying however at a given sentiment polarity. Moreover, the extraction this time round had a higher number of tweets and gives a considerable proportion of 72% for the positive tweets and a 28% for the negative. As a result, it showed how the various entities or concepts were directly linked to positivity or negativity of the sentiment. It analyzed that a good percentage as it is expected of positivity was correct as well as that of negativity. However, it incorporated some percentages of negative side to the positive and those of positive to negative. In consequence, it may be argued that, even though the second process is much demanding in acquiring tweets and then analyzing them, it is much accurate than the first process especially in the case involving huge volumes of data. Moreover, the gotten data could be analyzed deeper to according to either it being positive and negative. It could be analyzed into who it involved-the company, workers, sales staff and any other person involved. Moreover, it was noted that for the two applied types, they give out results related to features of the search. For the music or sports related, there is one where conversion of search results is higher. Part Two Natural language processing (NPL) is the part of computer studies that deals with artificial intelligence especially involving interactions of machine-computers-languages and the linguistic part of human. It basically tries to give out a meaning of human language to computers. Note that most of NPL are in machine language form and the now a days task is trying to remove the human task in coding directly by hand following statistical rules but to incorporate machine automation process. Therefore, machine will need to be enabled to learn about human slang, speech marks, and eventually be updating with language evolution. However, text mining is explained as the process of analyzing data contained in the natural language text. The act of obtaining data is usually referred to as text analysis and it has turned to be a useful tool to organization because they can mine useful information from postings in tweeter, linkedin or facebook. It takes NPL to mine unstructured data together with statistical techniques and machine techniques of learning. On the other hand, sentimental analysis can also be called opinion mining, appraisal extraction or subjectivity analysis, is the act of applying NPL, text mining, and computational methodologies in attempt to identify or extract information that is subjective from its source. It generally gives the writers of the text attitude in relation to the chosen topic. It mostly involves taking the sentiment holder, and from him, you attempt to identify with his target. In the process heavy grammatical relationships between words are established, and consequently polarity is determined as either negative or positive. These knowledge based utilities are however based on the available resources like WordNet. WordNet can easily be used in analyzing the sentimental analysis through it linguistic application extension called WordNet-effect. Through the extension, useful applications have resulted. For instance, this system classifies the texts into categories depending on emotion it is detected to carry. The system compares the in fed dictionary and the emotions described previously and analysis what the message may be containing. In case of a new word, it may be added in dictionary alongside it emotions for future references. Moreover, it lead to development of Eliot`s effective reasoned which is a collection of several programs that can tell of human emotions. Also, information or tutorial tools were made with it and the system is used to give a topic out of generation from the in fed natural languages. Additionally, it has led to development of sensing systems (Carlo & Allesadro, 2009). Some of the newest development in NPL includes unmanned cars. They can navigate street without the help of driver. They can sense through waves on how far the car is from the obstacles. In text mining there include website application to fetch and post its tweets from tweeter regularly without any person manning it. Other websites can predict the future. Moreover, other development include sensors of the voice from a person to tell whether lying or not. Additionally, PC are being developed that can sense and react to voices. With time, they will help in filling the gap between manual instruction of the machine and the machine taking its own initiative. In future, it will lead to machines that do not depend on human at all. The effective computing in today`s world will be eliminated by machine interpretation; machines that will be stronger than human in reasoning but very supportive. Reference Carlo S. & Allesadro V.(2009). WordNet-Affect: An effective extension of WordNet. Retrieved from: http://hnk.ffzg.hr/bibl/lrec2004/pdf/369.pdf followtweeter@api .Twitter API application page. Retrieve from: https://dev.twitter.com/apps Seif H., Hassan, He Y., Yulani, Alani H., & Harith (2012). Semantic sentiment analysis of tweeter. Boston, USA. Retrieved from: http://iswc2012.semanticweb.org/ Read More