Please enable JavaScript.
Coggle requires JavaScript to display documents.
BD3 (basic steps (noise cleaning: remove means not needed for analysis (…
BD3
basic steps
noise cleaning: remove means not needed for analysis ( URLS, Hashtags, Usernames)
-
-
-
-
word level analysis
-
-
-
accurate classification in: positive, negative, neutral
-
Regression
degree to which the variation in variable x is related / can be explained by the variation i another variable y
linear correlation --> write equation, square line, predict values of y for values of x falling within the range of data
-
-
two types of data
-
categorial (represents number of items that have a feature, conform to Zipf distribution
-
-
structural breaks
new patterns, new predictive pattern
outliers, in less than 10 %
Clustering:
distances, cluster nearest
-
-
-
-
-
correlation = quantitative relationship between two interval or ratio level variables, multivariate when more thatn two independent variables --> -1 strong negative correlation
-
modelling: formula explaining behavior of system, descriptive lanuage --> distribution of data, prediction how variables change over time