Detecting contrast patterns in newspaper articles by combining discourse analysis and text mining

Abstract

Text mining aims at constructing classification models and finding interesting patterns in large text collections. This paper investigates the utility of applying these techniques to media analysis, more specifically to support discourse analysis of news reports about the 2007 Kenyan elections and post-election crisis in local (Kenyan) and Western (British and US) newspapers. It illustrates how text mining methods can assist discourse analysis by finding contrast patterns which provide evidence for ideological differences between local and international press coverage. Our experiments indicate that most significant differences pertain to the interpretive frame of the news events: whereas the newspapers from the UK and the US focus on ethnicity in their coverage, the Kenyan press concentrates on sociopolitical aspects.

Investigating the Relevant Agro Food Keyword in Malaysian Online Newspapers

“…Out of the total identified keyword's, they were repetitively found 720 times in all newspapers -The Sun (35.14%), News Straits Times (29.03%), and The Star Online (35.83%). The Sun #253 fish (24), palm oil (22), rice (19), fruits (18), livestock (16), vegetables (16), chicken (14), crops (12), beef (8), durian (8), farmers (8), paddy (8), padi (7), seafood (7), breeders (5), coffee (5), poultry (5), cattle (4), farms (4), meat (4), seeds (4), food (3), herbs (3), plants (3), animals (1), beverages (1), biodiesel (1), bird (1), citrus (1), cockle (1), crude (1), elephant (1), fields (1), frozen (1), lamb (1), mangos teen (1), mussels (1), mutton (1), onion (1), orange (1), orchard (1), park (1), pork's (1), prawn (1), roselle (1), sea (1), seedlings (1), shellfish (1), turtle (1), wildlife (1) News Straits Times #209 palm oil (36), fruits (16), fish (15), rice (14), vegetable (14), livestock's (12), plants (8), durian (7), padi (7), animals (7), meat (4), crop (4), animal (4), honey (4), rubber (4), coconut (4), spices (4), sea (4), chicken (4), plantations (3), beef (3), milk (3), farming (3), poultry (3), pudding (1), oyster (1), pelagic (1), sharks (1), dairy (1), grass (1), juice (1), crude (1), pure (1), deer (1), goats (1), seedlings (1), boneless (1), tropical (1), stingless (1), green (1), dorsata (1), super fruits (1), lemon (1), fauna (1), drink (1), salmon (1) The Star Online #258 Palm oil (37), Rice (23), Vegetables (21), Fruits (20), Livestock (16), Fish (13), Crops (9), Meat (9), Farming (8), padi (8), corn (7), poultry (7), animal (6), food (6), beef (5), breeders (4), cattle (4, chicken (4), eggs (4), seafood (4), seeds (4), fishmeal (3), organic (3), plantation (3), planting (3), arowana (1, beverage (1), birds (1), black (1), buffalo (1), coconut (1), cooking (1), cow (1), cultivation (1), fishing (1), gliders (1), granaries (1), grass (1), lizard (1), peelers (1), pepper (1), prawn fishing (1), programmes (1), rear (1), seaweed (1), spices (1), staple (1), stockpiles (1, sweet (1), tea (1), tropical (1), wild (1) From the observation ...…”

“…The table sorts the Agro relevant keywords based on the descending total frequencies. TABLE IV FREQUENCY MATRIC BETWEEN THE STAR ONLINE, NEWS STRAITS TIME, AND THE The information in Table IV shows that there are 12 Agro keywords (item no [1][2][3][4][5][6][7][8][9][10][11][12] which are considered as the most important as their appearance in all news providers-palm oil, rice, fruits, fish, vegetable, livestock, paddy, crop, chicken, animal, meat, and beef. The bar chart in Fig.6 shows the frequency (%) of top 12 Agro keywords found in The Star Online, The Sun Daily, and News Straits Time.…”

“…For example, in stock price forecasting, the information form mass media is used as a basis to forecast inter-day stock prices. The behaviors identified from the media is used to determine the stock's future market condition whether it will rise or fall [11] and utilized news mining in analyzing the 2007 Kenyan elections news report [12]. In the studies, the political information from the local newspapers was analyzed and compared with the news published in Western newspapers (British and US).…”

Discursive Construction of Political Categories and Moral Fields: God Versus Rights and Access in a Reproductive Health Legislative Debate

“…For instance, computational data‐mining techniques like text mining are seen to address methodological limitations of traditional qualitative coding and analysis especially when dealing with large collections of textual data (Evans, ). In the case of news discourse analysis, text mining has been argued to lend quantitative support and rigor to in‐depth qualitative interpretations of ideological differences found within national and local media coverage (Pollak, Coesemans, Daelemans, & Lavrač, ; see also Koller & Mautner ; Montiel, Salvador, See, & de Leon, ; O'Halloran ). Additionally, text mining presents compatibility with the concept and process of categorization.…”

“…Additionally, text mining presents compatibility with the concept and process of categorization. As a variant of data mining, text mining is able to detect both the highly salient and distinctive terms, including collocations of these words that unveil inconspicuous yet interesting meaning patterns that uniquely differentiate sets of documents (Pollak et al, ), such as legislative speeches belonging to opposing sides of the RH bill debate.…”