Letter dataset: Results of Text mining project 2020
Your name and UCO
PA164 and KD Lab FI MU Brno
Abstract. This work addresses the problem of ...
1 Data set and task description
1.1 Data set
link to the data set, number of instances, features, missing values etc , data preparation (e.g.
transformation into form)
1.2 Data set reduction
Feature reduction
introduce the learning curve, explain what number of feature you will use and why
Other reduction
voluntary; only in the case that you need eg. limit the number of instances
2 Description of the method used
2.1 General
what classifiers (CLF) and what outlier detection (OD) methods you used. Do not describe
them.
2.2 Document-term matrix
what representation (pre-processing, PP) you used (very likely binary, TF, TF/IDF)
3 Results
3.1 Overview
general description of your results
2
3.2 Comparison of different combinations of (PP+OD+CLF)
including graphs and statistical tests.
Which combination was the best in term of accuracy.
Is that combination PP+OD+CLF significantly better that the others? https://is.muni.cz/auth/el/fi/podzim
3.3 Discussion
including discussion of the outliers that you found interesting
4 Conclusion
5 References
only if you employed something uncommon.