7. 附錄
7.2. 一般文字稿最低 Perplexity 值之主題數
未刪章節未加入提示詞時各摘要方法與一般 LDA 摘要方法和 PLSA 摘要方 法比較結果如表 46 所示,與動詞 LDA 摘要方法和名詞 LDA 摘要方法比較結果 如表 47 所示,與線上摘要器 SweSum 比較結果如表 48 所示。
表 46. 與 PLSA 摘要方法和一般 LDA 摘要方法比較結果
119
課程名稱 一般
LDA
TF TF-ISF TF-SF PLSA The Hardware/Software Interface 0.3126 0.3252 0.3257 0.321 0.3079 Social Network Analysis 0.3405 0.3589 0.3606 0.3492 0.3296 Web Application Architectures 0.4236 0.4562 0.4484 0.4522 0.4304 Audio Signal Processing for Music
Applications
0.1734 0.1982 0.1975 0.1927 0.1791 Malicious Software and its
Underground Economy Two Sides to Every Story
0.3572 0.3703 0.368 0.3696 0.3409
Big Data Science with the BD2K-LINCS Data Coordination and Integration Center
0.3295 0.3449 0.3493 0.3363 0.331
Experimental Methods in Systems Biology
0.3241 0.3391 0.3461 0.3363 0.3332 Dynamical Modeling Methods for
Systems Biology
0.3341 0.3543 0.36 0.3475 0.3347 The Brain and Space 0.2833 0.3111 0.3106 0.3083 0.2978 Network Analysis in Systems
Biology
0.3122 0.3279 0.3287 0.3276 0.3227
automata 0.5121 0.5252 0.5312 0.5197 0.5083
Natural Language Processing 2013 0.2717 0.2855 0.2825 0.2825 0.2808 Beginning Game Programming with
C
0.2788 0.3215 0.3286 0.3178 0.2765 Climate Change 0.2987 0.3254 0.3265 0.3283 0.2851 Journalism Skills for Engaged
Citizens
0.3474 0.3519 0.351 0.3569 0.351 Algorithms Design and Analysis Part
1
0.284 0.3146 0.3108 0.3152 0.2876 Algorithms Design and Analysis Part
2
0.3198 0.3434 0.3404 0.3426 0.3265 Introduction to Chemistry Reactions
and Ratios
0.2704 0.2935 0.2985 0.2902 0.2726 Genomic and Precision Medicine 0.2874 0.2923 0.294 0.2877 0.2866 Epigenetic Control of Gene
Expression
0.338 0.3693 0.3662 0.3671 0.33 Take the Lead on Healthcare Quality
Improvement
0.3901 0.4084 0.4028 0.4037 0.3895 Surviving Disruptive Technologies 0.4158 0.4057 0.4016 0.4062 0.4041 Caries Management by Risk
Assessment CAMBRA
0.3657 0.4162 0.419 0.4026 0.3429 Computational Neuroscience 0.3143 0.3211 0.3176 0.322 0.3084 Introduction to Data Science 0.2598 0.2694 0.2706 0.2691 0.2571 Discrete Optimization 0.2928 0.3 0.3047 0.3007 0.2846 Foundations of Virtual Instruction 0.2769 0.292 0.2838 0.2877 0.232 Virology I How Viruses Work 0.2429 0.2597 0.2593 0.2561 0.238 Edx Introduction to Computer
Science
120 with Java Part 1 Starting to Code
with Java
edx Programming Basics 0.3789 0.3742 0.3773 0.3741 0.3789 總平均 0.3155 0.3321 0.332 0.3296 0.3126
平均 F1 值:與一般 LDA 摘要方法呈現顯著差異
平均 F1 值:與 PLSA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比一般 LDA 摘要方法或 PLSA 摘要方法高
在表 46 的結果中除了 Surviving Disruptive Technologies 課程、edx Cellular mechanisms of brain function 課程和 edx Programming Basics 課程的 TF 方法、
The Hardware/SoftwareInterface
0.3158 0.3252 0.3257 0.321 0.313 Social Network Analysis 0.3466 0.3589 0.3606 0.3492 0.3503 Web Application Architectures 0.4286 0.4562 0.4484 0.4522 0.4399 Audio Signal Processing for
Music Applications
0.1879 0.1982 0.1975 0.1927 0.1734 Malicious Software and its
Underground Economy Two Sides to Every Story
0.3101 0.3703 0.368 0.3696 0.31
Big Data Science with the
BD2K-LINCS Data
Coordination and Integration Center
0.3408 0.3449 0.3493 0.3363 0.3425
Experimental Methods in Systems Biology
0.3268 0.3391 0.3461 0.3363 0.3313 Dynamical Modeling Methods
for Systems Biology
0.337 0.3543 0.36 0.3475 0.3476 The Brain and Space 0.2957 0.3111 0.3106 0.3083 0.2933 Network Analysis in Systems
Biology
0.3174 0.3279 0.3287 0.3276 0.3173
automata 0.5086 0.5252 0.5312 0.5197 0.5134
Natural Language Processing 2013
0.2711 0.2855 0.2825 0.2825 0 Beginning Game Programming
with C
0.2924 0.3215 0.3286 0.3178 0.2988 Climate Change 0.312 0.3254 0.3265 0.3283 0.3028
121 Journalism Skills for Engaged
Citizens
0.3484 0.3519 0.351 0.3569 0.3554 Algorithms Design and Analysis
Part 1
0.2863 0.3146 0.3108 0.3152 0.2879 Algorithms Design and Analysis
Part 2
0.32 0.3434 0.3404 0.3426 0.3166 Introduction to Chemistry
Reactions and Ratios
0.2826 0.2935 0.2985 0.2902 0.277 Genomic and Precision
Medicine Surviving Disruptive
Technologies
0.4096 0.4057 0.4016 0.4062 0.4131 Caries Management by Risk
Assessment CAMBRA
0.3761 0.4162 0.419 0.4026 0.3604 Computational Neuroscience 0.3194 0.3211 0.3176 0.322 0.3044 Introduction to Data Science 0.2593 0.2694 0.2706 0.2691 0.2646 Discrete Optimization 0.2894 0.3 0.3047 0.3007 0.2938 Foundations of Virtual
Instruction
0.2665 0.292 0.2838 0.2877 0.2477 Virology I How Viruses Work 0.2527 0.2597 0.2593 0.2561 0.2392 Edx Introduction to Computer
Science Starting to Code with Java
0.1413 0.1386 0.1299 0.1424 0.1401
edx Programming Basics 0.3799 0.3742 0.3773 0.3741 0.3764 總平均 0.3176 0.3321 0.332 0.3296 0.3081
平均 F1 值:與動詞 LDA 摘要方法呈現顯著差異
平均 F1 值:與名詞 LDA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比動詞 LDA 摘要方法或名詞 LDA 摘要方法高
在表 47 的結果中除了 Surviving Disruptive Technologies 課程、edx Cellular mechanisms of brain function 課程和 edx Programming Basics 課程的 TF 方法、
TF-ISF 方法和 TF-SF 方法平均 F1 值皆比動詞 LDA 摘要方法或名詞 LDA 摘要方
法低,其餘課程的 TF 方法、TF-ISF 方法和 TF-SF 方法平均 F1 值至少有一個摘
要方法平比動詞 LDA 摘要方法和名詞 LDA 摘要方法高。而在 Natural Language
Processing 2013 課程中因為其章節數過多而無法將名詞有效的分配置全部主題,
122
因此在本論文不另外建立主題模型。
表 48. 與線上摘要器 SweSum 比較結果
課程名稱 TF TF-ISF TF-SF SweSum
The Hardware/Software Interface 0.3252 0.3257 0.321 0.3189 Social Network Analysis 0.3589 0.3606 0.3492 0.358 Web Application Architectures 0.4562 0.4484 0.4522 0.4658 Audio Signal Processing for Music Applications 0.1982 0.1975 0.1927 0.2035 Malicious Software and its Underground
Economy Two Sides to Every Story
0.3703 0.368 0.3696 0.3637 Big Data Science with the BD2K-LINCS Data
Coordination and Integration Center
0.3449 0.3493 0.3363 0.3322 Experimental Methods in Systems Biology 0.3391 0.3461 0.3363 0.3358 Dynamical Modeling Methods for Systems
Biology
0.3543 0.36 0.3475 0.3408 The Brain and Space 0.3111 0.3106 0.3083 0.3096 Network Analysis in Systems Biology 0.3279 0.3287 0.3276 0.3203
automata 0.5252 0.5312 0.5197 0.515
Natural Language Processing 2013 0.2855 0.2825 0.2825 0.2538 Beginning Game Programming with C 0.3215 0.3286 0.3178 0.3069
Climate Change 0.3254 0.3265 0.3283 0.3214
Journalism Skills for Engaged Citizens 0.3519 0.351 0.3569 0.352 Algorithms Design and Analysis Part 1 0.3146 0.3108 0.3152 0.2971 Algorithms Design and Analysis Part 2 0.3434 0.3404 0.3426 0.3319 Introduction to Chemistry Reactions and Ratios 0.2935 0.2985 0.2902 0.2849 Genomic and Precision Medicine 0.2923 0.294 0.2877 0.2855 Epigenetic Control of Gene Expression 0.3693 0.3662 0.3671 0.3605 Take the Lead on Healthcare Quality
Improvement
0.4084 0.4028 0.4037 0.3891 Surviving Disruptive Technologies 0.4057 0.4016 0.4062 0.4287 Caries Management by Risk Assessment
CAMBRA
0.4162 0.419 0.4026 0.4219 Computational Neuroscience 0.3211 0.3176 0.322 0.3337 Introduction to Data Science 0.2694 0.2706 0.2691 0.2667
Discrete Optimization 0.3 0.3047 0.3007 0.2993
Foundations of Virtual Instruction 0.292 0.2838 0.2877 0.255 Virology I How Viruses Work 0.2597 0.2593 0.2561 0.2487 Edx Introduction to Computer Science 0.3092 0.3064 0.3089 0.292 edx Big Data in Education 0.4794 0.4826 0.4773 0.4781 edx Cellular mechanisms of brain function 0.1765 0.1768 0.1765 0.176 edx Introduction to Programming with Java Part
1 Starting to Code with Java
0.1386 0.1299 0.1424 0.151 edx Programming Basics 0.3742 0.3773 0.3741 0.3674
總平均 0.3321 0.332 0.3296 0.3262
平均 F1 值:與線上摘要器 SweSum 呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比線上摘要器 SweSum 高
123
在表 48 的結果中除了 Web Application Architectures 課程、Audio Signal Processing for Music Applications 課程、Surviving Disruptive Technologies 課程、
Caries Management by Risk Assessment CAMBRA 課 程 、 Computational Neuroscience 課程和 edx Introduction to Programming with Java Part 1 Starting to Code with Java 課程的 TF 方法、TF-ISF 方法和 TF-SF 方法平均 F1 值皆比線上摘 要器 SweSum 低,其餘課程的 TF 方法、TF-ISF 方法和 TF-SF 方法平均 F1 值至 少有一個摘要方法比線上摘要器 SweSum 高。
未刪章節使用提示詞 TF 權重正規化加權提示詞時各摘要方法與 PLSA 摘要 方法和一般 LDA 摘要方法比較結果如表 49 所示,與名詞 LDA 摘要方法和動詞 LDA 摘要方法比較結果如表 50 所示,與線上摘要器 SweSum 比較結果如表 51 所示。
表 49. 與 PLSA 摘要方法和一般 LDA 摘要方法比較結果
課程名稱 一般
LDA
TF TF-ISF TF-SF PLSA The Hardware/Software Interface 0.3126 0.3202 0.3266 0.3088 0.3079 Social Network Analysis 0.3405 0.3443 0.3539 0.3343 0.3296 Web Application Architectures 0.4236 0.4352 0.4408 0.4295 0.4304 Audio Signal Processing for Music
Applications
0.1734 0.1926 0.1965 0.1853 0.1791 Malicious Software and its
Underground Economy Two Sides to Every Story
0.3572 0.3512 0.3601 0.3335 0.3409
Big Data Science with the BD2K-LINCS Data Coordination and Integration Center
0.3295 0.3323 0.339 0.3255 0.331
Experimental Methods in Systems Biology
0.3241 0.3338 0.3403 0.3223 0.3332 Dynamical Modeling Methods for
Systems Biology
0.3341 0.3481 0.3552 0.3319 0.3347 The Brain and Space 0.2833 0.2978 0.3119 0.2829 0.2978 Network Analysis in Systems
Biology
0.3122 0.3253 0.3288 0.3177 0.3227
automata 0.5121 0.5302 0.5307 0.5274 0.5083
Natural Language Processing 2013 0.2717 0.2759 0.2766 0.267 0.2808 Beginning Game Programming with
C
0.2788 0.3121 0.3208 0.2927 0.2765 Climate Change 0.2987 0.3253 0.3222 0.3061 0.2851
124 Journalism Skills for Engaged
Citizens
0.3474 0.3445 0.3446 0.3379 0.351 Algorithms Design and Analysis Part
1
0.284 0.307 0.3114 0.2968 0.2876 Algorithms Design and Analysis Part
2
0.3198 0.3395 0.3407 0.3354 0.3265 Introduction to Chemistry Reactions
and Ratios
0.2704 0.2855 0.2941 0.2631 0.2726 Genomic and Precision Medicine 0.2874 0.287 0.2936 0.2823 0.2866 Epigenetic Control of Gene
Expression
0.338 0.3583 0.3618 0.3353 0.33 Take the Lead on Healthcare Quality
Improvement
0.3901 0.3896 0.3954 0.3776 0.3895 Surviving Disruptive Technologies 0.4158 0.3876 0.3959 0.3805 0.4041 Caries Management by Risk
Assessment CAMBRA
0.3657 0.3812 0.4052 0.3504 0.3429 Computational Neuroscience 0.3143 0.3161 0.3205 0.3048 0.3084 Introduction to Data Science 0.2598 0.2585 0.2665 0.2524 0.2571 Discrete Optimization 0.2928 0.2954 0.301 0.2781 0.2846 Foundations of Virtual Instruction 0.2769 0.2469 0.2781 0.221 0.232 Virology I How Viruses Work 0.2429 0.2439 0.2462 0.2317 0.238 Edx Introduction to Computer
Science
0.2967 0.2997 0.3033 0.2923 0.2872 edx Big Data in Education 0.4708 0.4731 0.4826 0.462 0.4682 edx Cellular mechanisms of brain
function
0.1777 0.1756 0.1768 0.1768 0.1788 edx Introduction to Programming
with Java Part 1 Starting to Code with Java
0.1292 0.1335 0.1279 0.1335 0.134
edx Programming Basics 0.3789 0.3715 0.3762 0.3725 0.3789 總平均 0.3155 0.3218 0.328 0.3106 0.3126
平均 F1 值:與一般 LDA 摘要方法呈現顯著差異
平均 F1 值:與 PLSA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比一般 LDA 摘要方法或 PLSA 摘要方法高
在表 49 的結果中 Natural Language Processing 2013 課程、Journalism Skills for Engaged Citizens 課程、Surviving Disruptive Technologies 課程、edx Cellular mechanisms of brain function 課程、edx Introduction to Programming with Java Part 1 Starting to Code with Java 課程和 edx Programming Basics 課程的 TF 摘要方法、
TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值皆比 PLSA 摘要方法和一般 LDA
摘要方法低,其餘課程的 TF 方法、TF-ISF 方法和 TF-SF 方法平均 F1 值至少有
一個摘要方法比 PLSA 摘要方法和一般 LDA 摘要方法高。
125 The Hardware/Software Interface 0.3158 0.3202 0.3266 0.3088 0.313 Social Network Analysis 0.3466 0.3443 0.3539 0.3343 0.3503 Web Application Architectures 0.4286 0.4352 0.4408 0.4295 0.4399 Audio Signal Processing for Music
Applications
0.1879 0.1926 0.1965 0.1853 0.1734 Malicious Software and its
Underground Economy Two Sides to Every Story
0.3101 0.3512 0.3601 0.3335 0.31
Big Data Science with the BD2K-LINCS Data Coordination and Integration Center
0.3408 0.3323 0.339 0.3255 0.3425
Experimental Methods in Systems Biology
0.3268 0.3338 0.3403 0.3223 0.3313 Dynamical Modeling Methods for
Systems Biology
0.337 0.3481 0.3552 0.3319 0.3476 The Brain and Space 0.2957 0.2978 0.3119 0.2829 0.2933 Network Analysis in Systems
Biology
0.3174 0.3253 0.3288 0.3177 0.3173
automata 0.5086 0.5302 0.5307 0.5274 0.5134
Natural Language Processing 2013 0.2711 0.2759 0.2766 0.267 0 Beginning Game Programming
with C
0.2924 0.3121 0.3208 0.2927 0.2988 Climate Change 0.312 0.3253 0.3222 0.3061 0.3028 Journalism Skills for Engaged
Citizens
0.3484 0.3445 0.3446 0.3379 0.3554 Algorithms Design and Analysis
Part 1
0.2863 0.307 0.3114 0.2968 0.2879 Algorithms Design and Analysis
Part 2
0.32 0.3395 0.3407 0.3354 0.3166 Introduction to Chemistry
Reactions and Ratios
0.2826 0.2855 0.2941 0.2631 0.277 Genomic and Precision Medicine 0.2891 0.287 0.2936 0.2823 0.2858 Epigenetic Control of Gene Surviving Disruptive Technologies 0.4096 0.3876 0.3959 0.3805 0.4131 Caries Management by Risk
Assessment CAMBRA
0.3761 0.3812 0.4052 0.3504 0.3604 Computational Neuroscience 0.3194 0.3161 0.3205 0.3048 0.3044 Introduction to Data Science 0.2593 0.2585 0.2665 0.2524 0.2646 Discrete Optimization 0.2894 0.2954 0.301 0.2781 0.2938 Foundations of Virtual Instruction 0.2665 0.2469 0.2781 0.221 0.2477 Virology I How Viruses Work 0.2527 0.2439 0.2462 0.2317 0.2392 Edx Introduction to Computer
Science
0.2925 0.2997 0.3033 0.2923 0.2926 edx Big Data in Education 0.4669 0.4731 0.4826 0.462 0.4727 edx Cellular mechanisms of brain 0.1805 0.1756 0.1768 0.1768 0.1786
126 function
edx Introduction to Programming with Java Part 1 Starting to Code with Java
0.1413 0.1335 0.1279 0.1335 0.1401
edx Programming Basics 0.3799 0.3715 0.3762 0.3725 0.3764 總平均 0.3176 0.3218 0.328 0.3106 0.3081
平均 F1 值:與動詞 LDA 摘要方法呈現顯著差異
平均 F1 值:與名詞 LDA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比動詞 LDA 摘要方法或名詞 LDA 摘要方法高
在表 50 的結果中 Big Data Science with the BD2K-LINCS Data Coordination and Integration Center 課程、Journalism Skills for Engaged Citizens 課程、Surviving Disruptive Technologies 課程、Virology I How Viruses Work 課程、edx Cellular mechanisms of brain function 課程、edx Introduction to Programming with Java Part 1 Starting to Code with Java 課程和 edx Programming Basics 課程的 TF 摘要方法、
TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值皆比名詞 LDA 摘要方法或動詞 LDA 摘要方法低,其餘課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方 法平均 F1 值至少有一個摘要方法比名詞 LDA 摘要方法和動詞 LDA 摘要方法 高。
表 51. 與線上摘要器 SweSum 比較結果
課程名稱 TF TF-ISF TF-SF SweSum
The Hardware/Software Interface 0.3202 0.3266 0.3088 0.3189 Social Network Analysis 0.3443 0.3539 0.3343 0.358 Web Application Architectures 0.4352 0.4408 0.4295 0.4658 Audio Signal Processing for Music
Applications
0.1926 0.1965 0.1853 0.2035 Malicious Software and its Underground
Economy Two Sides to Every Story
0.3512 0.3601 0.3335 0.3637 Big Data Science with the BD2K-LINCS Data
Coordination and Integration Center
0.3323 0.339 0.3255 0.3322 Experimental Methods in Systems Biology 0.3338 0.3403 0.3223 0.3358 Dynamical Modeling Methods for Systems
Biology
0.3481 0.3552 0.3319 0.3408 The Brain and Space 0.2978 0.3119 0.2829 0.3096 Network Analysis in Systems Biology 0.3253 0.3288 0.3177 0.3203
automata 0.5302 0.5307 0.5274 0.515
Natural Language Processing 2013 0.2759 0.2766 0.267 0.2538 Beginning Game Programming with C 0.3121 0.3208 0.2927 0.3069
Climate Change 0.3253 0.3222 0.3061 0.3214
127
Journalism Skills for Engaged Citizens 0.3445 0.3446 0.3379 0.352 Algorithms Design and Analysis Part 1 0.307 0.3114 0.2968 0.2971 Algorithms Design and Analysis Part 2 0.3395 0.3407 0.3354 0.3319 Introduction to Chemistry Reactions and
Ratios
0.2855 0.2941 0.2631 0.2849 Genomic and Precision Medicine 0.287 0.2936 0.2823 0.2855 Epigenetic Control of Gene Expression 0.3583 0.3618 0.3353 0.3605 Take the Lead on Healthcare Quality
Improvement
0.3896 0.3954 0.3776 0.3891 Surviving Disruptive Technologies 0.3876 0.3959 0.3805 0.4287 Caries Management by Risk Assessment
CAMBRA
0.3812 0.4052 0.3504 0.4219 Computational Neuroscience 0.3161 0.3205 0.3048 0.3337 Introduction to Data Science 0.2585 0.2665 0.2524 0.2667 Discrete Optimization 0.2954 0.301 0.2781 0.2993 Foundations of Virtual Instruction 0.2469 0.2781 0.221 0.255 Virology I How Viruses Work 0.2439 0.2462 0.2317 0.2487 Edx Introduction to Computer Science 0.2997 0.3033 0.2923 0.292 edx Big Data in Education 0.4731 0.4826 0.462 0.4781 edx Cellular mechanisms of brain function 0.1756 0.1768 0.1768 0.176 edx Introduction to Programming with Java
Part 1 Starting to Code with Java
0.1335 0.1279 0.1335 0.151 edx Programming Basics 0.3715 0.3762 0.3725 0.3674
總平均 0.3218 0.328 0.3106 0.3262
平均 F1 值:與線上摘要器 SweSum 呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比線上摘要器 SweSum 高
在表 51 的結果中除了 Social Network Analysis 課程、Web Application Architectures 課程、Audio Signal Processing for Music Applications 課程、Malicious Software and its Underground Economy Two Sides to Every Story 課程、Surviving Disruptive Technologies 課程、Caries Management by Risk Assessment CAMBRA 課程、Computational Neuroscience 課程、Virology I How Viruses Work 課程和 edx Introduction to Programming with Java Part 1 Starting to Code with Java 課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值皆比線上摘要器 SweSum 低,其餘課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值至 少有一個摘要方法比線上摘要器 SweSum 高。
未刪章節使用提示詞出現次數乘上提示詞 TF 權重加權提示詞時各摘要方法
與 PLSA 摘要方法和一般 LDA 摘要方法比較結果如表 52 所示,與名詞 LDA 摘
128 The Hardware/Software Interface 0.3126 0.327 0.3284 0.3219 0.3079 Social Network Analysis 0.3405 0.3509 0.3519 0.3422 0.3296 Web Application Architectures 0.4236 0.4131 0.4168 0.4131 0.4304 Audio Signal Processing for Music
Applications
0.1734 0.1784 0.1802 0.1727 0.1791 Malicious Software and its Underground
Economy Two Sides to Every Story
0.3572 0.3333 0.3322 0.3261 0.3409 Big Data Science with the
BD2K-LINCS Data Coordination and Integration Center
0.3295 0.352 0.3579 0.3426 0.331
Experimental Methods in Systems Biology
0.3241 0.0903 0.0904 0.0886 0.3332 Dynamical Modeling Methods for
Systems Biology
0.3341 0.3838 0.3897 0.3732 0.3347 The Brain and Space 0.2833 0.3072 0.3087 0.3018 0.2978 Network Analysis in Systems Biology 0.3122 0.3355 0.3368 0.3325 0.3227
automata 0.5121 0.5076 0.514 0.5031 0.5083
Natural Language Processing 2013 0.2717 0.2801 0.2768 0.2757 0.2808 Beginning Game Programming with C 0.2788 0.3043 0.3079 0.3002 0.2765 Climate Change 0.2987 0.3317 0.3289 0.3328 0.2851 Journalism Skills for Engaged Citizens 0.3474 0.3533 0.3512 0.3572 0.351 Algorithms Design and Analysis Part 1 0.284 0.3132 0.3108 0.3152 0.2876 Algorithms Design and Analysis Part 2 0.3198 0.3424 0.3409 0.3425 0.3265 Introduction to Chemistry Reactions and
Ratios
0.2704 0.2282 0.2304 0.2219 0.2726 Genomic and Precision Medicine 0.2874 0.2921 0.2938 0.2868 0.2866 Epigenetic Control of Gene Expression 0.338 0.3694 0.366 0.3663 0.33 Take the Lead on Healthcare Quality
Improvement
0.3901 0.4068 0.4027 0.3993 0.3895 Surviving Disruptive Technologies 0.4158 0.4068 0.4042 0.4008 0.4041 Caries Management by Risk Assessment
CAMBRA
0.3657 0.4154 0.4173 0.396 0.3429 Computational Neuroscience 0.3143 0.3221 0.3176 0.3203 0.3084 Introduction to Data Science 0.2598 0.2691 0.2707 0.2666 0.2571 Discrete Optimization 0.2928 0.3007 0.3041 0.298 0.2846 Foundations of Virtual Instruction 0.2769 0.2833 0.2838 0.2802 0.232 Virology I How Viruses Work 0.2429 0.2472 0.2476 0.242 0.238 Edx Introduction to Computer Science 0.2967 0.3091 0.3062 0.3059 0.2872 edx Big Data in Education 0.4708 0.4789 0.4828 0.4773 0.4682 edx Cellular mechanisms of brain
function
0.1777 0.1768 0.1769 0.1777 0.1788 edx Introduction to Programming with 0.1292 0.1379 0.129 0.1414 0.134
129 Java Part 1 Starting to Code with Java
edx Programming Basics 0.3789 0.3736 0.377 0.3729 0.3789 總平均 0.3155 0.3188 0.3192 0.315 0.3126
平均 F1 值:與一般 LDA 摘要方法呈現顯著差異
平均 F1 值:與 PLSA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比一般 LDA 摘要方法或 PLSA 摘要方法高
在表 52 的結果中除了 Web Application Architectures 課程、Audio Signal Processing for Music Applications 課程、Malicious Software and its Underground Economy Two Sides to Every Story 課程、Experimental Methods in Systems Biology 課程、Introduction to Chemistry Reactions and Ratios 課程、Surviving Disruptive Technologies 課程和 edx Programming Basics 課程的 TF 摘要方法、TF-ISF 摘要方 法和 TF-SF 摘要方法平均 F1 值皆比一般 LDA 摘要方法或 PLSA 摘要方法低,
其餘課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值至少有 一個摘要方法比一般 LDA 摘要方法和 PLSA 摘要方法高。
表 53. 與名詞 LDA 摘要方法和動詞 LDA 摘要方法比較結果
課程名稱 動詞
LDA
TF TF-ISF TF-SF 名詞 LDA The Hardware/Software Interface 0.3158 0.327 0.3284 0.3219 0.313 Social Network Analysis 0.3466 0.3509 0.3519 0.3422 0.3503 Web Application Architectures 0.4286 0.4131 0.4168 0.4131 0.4399 Audio Signal Processing for Music
Applications
0.1879 0.1784 0.1802 0.1727 0.1734 Malicious Software and its Underground
Economy Two Sides to Every Story
0.3101 0.3333 0.3322 0.3261 0.31 Big Data Science with the
BD2K-LINCS Data Coordination and Integration Center
0.3408 0.352 0.3579 0.3426 0.3425
Experimental Methods in Systems Biology
0.3268 0.0903 0.0904 0.0886 0.3313 Dynamical Modeling Methods for
Systems Biology
0.337 0.3838 0.3897 0.3732 0.3476 The Brain and Space 0.2957 0.3072 0.3087 0.3018 0.2933 Network Analysis in Systems Biology 0.3174 0.3355 0.3368 0.3325 0.3173
automata 0.5086 0.5076 0.514 0.5031 0.5134
Natural Language Processing 2013 0.2711 0.2801 0.2768 0.2757 0 Beginning Game Programming with C 0.2924 0.3043 0.3079 0.3002 0.2988 Climate Change 0.312 0.3317 0.3289 0.3328 0.3028 Journalism Skills for Engaged Citizens 0.3484 0.3533 0.3512 0.3572 0.3554
130
Algorithms Design and Analysis Part 1 0.2863 0.3132 0.3108 0.3152 0.2879 Algorithms Design and Analysis Part 2 0.32 0.3424 0.3409 0.3425 0.3166 Introduction to Chemistry Reactions and
Ratios
0.2826 0.2282 0.2304 0.2219 0.277 Genomic and Precision Medicine 0.2891 0.2921 0.2938 0.2868 0.2858 Epigenetic Control of Gene Expression 0.3367 0.3694 0.366 0.3663 0.335 Take the Lead on Healthcare Quality
Improvement
0.3918 0.4068 0.4027 0.3993 0.393 Surviving Disruptive Technologies 0.4096 0.4068 0.4042 0.4008 0.4131 Caries Management by Risk Assessment
CAMBRA
0.3761 0.4154 0.4173 0.396 0.3604 Computational Neuroscience 0.3194 0.3221 0.3176 0.3203 0.3044 Introduction to Data Science 0.2593 0.2691 0.2707 0.2666 0.2646 Discrete Optimization 0.2894 0.3007 0.3041 0.298 0.2938 Foundations of Virtual Instruction 0.2665 0.2833 0.2838 0.2802 0.2477 Virology I How Viruses Work 0.2527 0.2472 0.2476 0.242 0.2392 Edx Introduction to Computer Science 0.2925 0.3091 0.3062 0.3059 0.2926 edx Big Data in Education 0.4669 0.4789 0.4828 0.4773 0.4727 edx Cellular mechanisms of brain
function
0.1805 0.1768 0.1769 0.1777 0.1786 edx Introduction to Programming with
Java Part 1 Starting to Code with Java
0.1413 0.1379 0.129 0.1414 0.1401 edx Programming Basics 0.3799 0.3736 0.377 0.3729 0.3764 總平均 0.3176 0.3188 0.3192 0.315 0.3081
平均 F1 值:與動詞 LDA 摘要方法呈現顯著差異
平均 F1 值:與名詞 LDA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比動詞 LDA 摘要方法或名詞 LDA 摘要方法高
在表 53 的結果中除了 Web Application Architectures 課程、Audio Signal Processing for Music Applications 課程、Malicious Software and its Underground Economy Two Sides to Every Story 課程、Experimental Methods in Systems Biology 課程、Introduction to Chemistry Reactions and Ratios 課程、Surviving Disruptive Technologies 課程和 edx Programming Basics 課程的 TF 摘要方法、TF-ISF 摘要方 法和 TF-SF 摘要方法平均 F1 值皆比動詞 LDA 摘要方法或名詞 LDA摘要方法低,
其餘課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值至少有 一個摘要方法比動詞 LDA 摘要方法和名詞 LDA 摘要方法高。
表 54. 與線上摘要器 SweSum 比較結果
課程名稱 TF TF-ISF TF-SF SweSum
The Hardware/Software Interface 0.327 0.3284 0.3219 0.3189
131
Social Network Analysis 0.3509 0.3519 0.3422 0.358 Web Application Architectures 0.4131 0.4168 0.4131 0.4658 Audio Signal Processing for Music Applications 0.1784 0.1802 0.1727 0.2035 Malicious Software and its Underground
Economy Two Sides to Every Story
0.3333 0.3322 0.3261 0.3637 Big Data Science with the BD2K-LINCS Data
Coordination and Integration Center
0.352 0.3579 0.3426 0.3322 Experimental Methods in Systems Biology 0.0903 0.0904 0.0886 0.3358 Dynamical Modeling Methods for Systems
Biology
0.3838 0.3897 0.3732 0.3408 The Brain and Space 0.3072 0.3087 0.3018 0.3096 Network Analysis in Systems Biology 0.3355 0.3368 0.3325 0.3203
automata 0.5076 0.514 0.5031 0.515
Natural Language Processing 2013 0.2801 0.2768 0.2757 0.2538 Beginning Game Programming with C 0.3043 0.3079 0.3002 0.3069
Climate Change 0.3317 0.3289 0.3328 0.3214
Journalism Skills for Engaged Citizens 0.3533 0.3512 0.3572 0.352 Algorithms Design and Analysis Part 1 0.3132 0.3108 0.3152 0.2971 Algorithms Design and Analysis Part 2 0.3424 0.3409 0.3425 0.3319 Introduction to Chemistry Reactions and Ratios 0.2282 0.2304 0.2219 0.2849 Genomic and Precision Medicine 0.2921 0.2938 0.2868 0.2855 Epigenetic Control of Gene Expression 0.3694 0.366 0.3663 0.3605 Take the Lead on Healthcare Quality
Improvement
0.4068 0.4027 0.3993 0.3891 Surviving Disruptive Technologies 0.4068 0.4042 0.4008 0.4287 Caries Management by Risk Assessment
CAMBRA
0.4154 0.4173 0.396 0.4219 Computational Neuroscience 0.3221 0.3176 0.3203 0.3337 Introduction to Data Science 0.2691 0.2707 0.2666 0.2667 Discrete Optimization 0.3007 0.3041 0.298 0.2993 Foundations of Virtual Instruction 0.2833 0.2838 0.2802 0.255 Virology I How Viruses Work 0.2472 0.2476 0.242 0.2487 Edx Introduction to Computer Science 0.3091 0.3062 0.3059 0.292 edx Big Data in Education 0.4789 0.4828 0.4773 0.4781 edx Cellular mechanisms of brain function 0.1768 0.1769 0.1777 0.176 edx Introduction to Programming with Java Part
1 Starting to Code with Java
0.1379 0.129 0.1414 0.151 edx Programming Basics 0.3736 0.377 0.3729 0.3674
總平均 0.3188 0.3192 0.315 0.3262
平均 F1 值:與線上摘要器 SweSum 呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比線上摘要器 SweSum 高
在表 54 的結果中除了 Social Network Analysis 課程、Web Application
Architectures 課程、Audio Signal Processing for Music Applications 課程、Malicious
Software and its Underground Economy Two Sides to Every Story 課 程 、
Experimental Methods in Systems Biology 課程、The Brain and Space 課程、automata
132
課程、Introduction to Chemistry Reactions and Ratios 課程、Surviving Disruptive Technologies 課程、Caries Management by Risk Assessment CAMBRA 課程、
Computational Neuroscience 課程、Virology I How Viruses Work 課程和 edx Introduction to Programming with Java Part 1 Starting to Code with Java 課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值皆比線上摘要器 SweSum 低,其餘課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值至 少有一個摘要方法比線上摘要器 SweSum 高。
刪章節未加入提示詞時各摘要方法與 PLSA 摘要方法和一般 LDA 摘要方法 比較結果如表 55 所示,與名詞 LDA 摘要方法和動詞 LDA 摘要方法比較結果如 表 56 所示,與線上摘要器 SweSum 比較結果如表 57 所示。
表 55. 與 PLSA 摘要方法和一般 LDA 摘要方法比較結果
課程名稱 一般
LDA
TF TF-ISF TF-SF PLSA The Hardware/Software Interface 0.3123 0.3286 0.3306 0.324 0.3151 Social Network Analysis 0.3543 0.3601 0.3583 0.3501 0.3322 Web Application Architectures 0.4432 0.4532 0.4507 0.4518 0.4318 Audio Signal Processing for Music
Applications
0.1734 0.1982 0.1975 0.1927 0.1791 Malicious Software and its
Underground Economy Two Sides to Every Story
0.3539 0.3719 0.3694 0.3694 0.3361
Big Data Science with the BD2K-LINCS Data Coordination and Integration Center
0.2981 0.3119 0.3131 0.3021 0.2849
Experimental Methods in Systems Biology
0.3241 0.3391 0.3461 0.3364 0.3332 Dynamical Modeling Methods for
Systems Biology
0.3412 0.3577 0.364 0.3499 0.3309 The Brain and Space 0.2833 0.3111 0.3106 0.3083 0.2978 Network Analysis in Systems
Biology
0.3145 0.3183 0.3216 0.32 0.3151
automata 0.5121 0.5252 0.5312 0.5197 0.5083
Natural Language Processing 2013 0.2678 0.281 0.2744 0.2787 0.2664 Beginning Game Programming
with C
0.2788 0.3215 0.3286 0.3178 0.2765 Climate Change 0.2987 0.3254 0.3265 0.3283 0.2851 Journalism Skills for Engaged
Citizens
0.3491 0.3496 0.3482 0.3542 0.3264 Algorithms Design and Analysis 0.284 0.3147 0.3108 0.3152 0.2876
133 Part 1
Algorithms Design and Analysis Part 2
0.3198 0.3434 0.3404 0.3426 0.3265 Introduction to Chemistry
Reactions and Ratios
0.2841 0.3002 0.3037 0.2942 0.275 Genomic and Precision Medicine 0.2793 0.2923 0.294 0.2877 0.2672 Epigenetic Control of Gene Surviving Disruptive Technologies 0.4028 0.4141 0.4025 0.4145 0.3892 Caries Management by Risk
Assessment CAMBRA
0.3657 0.4162 0.419 0.4026 0.3429 Computational Neuroscience 0.3143 0.3211 0.3176 0.322 0.3084 Introduction to Data Science 0.2511 0.2688 0.2678 0.268 0.2563 Discrete Optimization 0.2928 0.3 0.3047 0.3007 0.2846 Foundations of Virtual Instruction 0.2769 0.292 0.2838 0.2877 0.232 Virology I How Viruses Work 0.2343 0.2461 0.2439 0.24 0.2282 Edx Introduction to Computer
Science edx Introduction to Programming
with Java Part 1 Starting to Code with Java
0.1262 0.1336 0.1318 0.1313 0.1324
edx Programming Basics 0.3772 0.3714 0.3729 0.3708 0.3794 總平均 0.3145 0.3303 0.3299 0.3273 0.3078
平均 F1 值:與一般 LDA 摘要方法呈現顯著差異
平均 F1 值:與 PLSA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比一般 LDA 摘要方法或 PLSA 摘要方法高
在表 55 的結果中除了 edx Cellular mechanisms of brain function 課程和 edx Programming Basics 課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平 均 F1 值皆比一般 LDA 摘要方法或 PLSA 摘要方法低,其餘課程的 TF 摘要方法、
The Hardware/SoftwareInterface
0.316 0.3286 0.3306 0.324 0.3171
134
Social Network Analysis 0.3465 0.3601 0.3583 0.3501 0.3461 Web Application Architectures 0.4334 0.4532 0.4507 0.4518 0.4337 Audio Signal Processing for
Music Applications
0.1879 0.1982 0.1975 0.1927 0.1734 Malicious Software and its
Underground Economy Two Sides to Every Story
0.3488 0.3719 0.3694 0.3694 0.3508
Big Data Science with the
BD2K-LINCS Data
Coordination and Integration Center
0.2933 0.3119 0.3131 0.3021 0.2988
Experimental Methods in Systems Biology
0.3268 0.3391 0.3461 0.3364 0.3314 Dynamical Modeling Methods
for Systems Biology
0.3347 0.3577 0.364 0.3499 0.3479 The Brain and Space 0.2957 0.3111 0.3106 0.3083 0.2933 Network Analysis in Systems
Biology
0.312 0.3183 0.3216 0.32 0.3174
automata 0.5086 0.5252 0.5312 0.5197 0.5134
Natural Language Processing 2013
0.2694 0.281 0.2744 0.2787 0 Beginning Game Programming
with C
0.2924 0.3215 0.3286 0.3178 0.2988 Climate Change 0.312 0.3254 0.3265 0.3283 0.3029 Journalism Skills for Engaged
Citizens
0.3411 0.3496 0.3482 0.3542 0.3332 Algorithms Design and Analysis
Part 1
0.2863 0.3147 0.3108 0.3152 0.2879 Algorithms Design and Analysis
Part 2
0.3201 0.3434 0.3404 0.3426 0.3166 Introduction to Chemistry
Reactions and Ratios
0.2773 0.3002 0.3037 0.2942 0.2783 Genomic and Precision
Medicine Surviving Disruptive
Technologies
0.4028 0.4141 0.4025 0.4145 0.407 Caries Management by Risk
Assessment CAMBRA
0.3761 0.4162 0.419 0.4026 0.3604 Computational Neuroscience 0.3194 0.3211 0.3176 0.322 0.3044 Introduction to Data Science 0.2506 0.2688 0.2678 0.268 0.2504 Discrete Optimization 0.2895 0.3 0.3047 0.3007 0.2938 Foundations of Virtual
Instruction
0.2665 0.292 0.2838 0.2877 0.2478 Virology I How Viruses Work 0.2343 0.2461 0.2439 0.24 0.2208 Edx Introduction to Computer
Science
0.2836 0.2993 0.298 0.3003 0.287 edx Big Data in Education 0.453 0.4736 0.4731 0.4659 0.4616 edx Cellular mechanisms of 0.1805 0.1765 0.1768 0.1765 0.1786
135 brain function
edx Introduction to Programming with Java Part 1 Starting to Code with Java
0.1223 0.1336 0.1318 0.1313 0.1321
edx Programming Basics 0.3702 0.3714 0.3729 0.3708 0.3777 總平均 0.3134 0.3303 0.3299 0.3273 0.3048
平均 F1 值:與動詞 LDA 摘要方法呈現顯著差異
平均 F1 值:與名詞 LDA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比動詞 LDA 摘要方法或名詞 LDA 摘要方法高
在表 56 的結果中除了 edx Cellular mechanisms of brain function 課程和 edx Programming Basics 課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平 均 F1 值皆比動詞 LDA 摘要方法或名詞 LDA 摘要方法低,其餘課程的 TF 摘要 方法、TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值至少有一個摘要方法比動詞 LDA 摘要方法和名 詞 LDA 摘 要 方法 摘要方 法 高 。而在 Natural Language Processing 2013 課程中因為其章節數過多而無法將名詞有效的分配置全部主題,
因此在本論文不另外建立主題模型。
表 57. 與線上摘要器 SweSum 比較結果
課程名稱 TF TF-ISF TF-SF SweSu
m The Hardware/Software Interface 0.3286 0.3306 0.324 0.321 Social Network Analysis 0.3601 0.3583 0.3501 0.3598 Web Application Architectures 0.4532 0.4507 0.4518 0.4658 Audio Signal Processing for Music
Applications
0.1982 0.1975 0.1927 0.2035 Malicious Software and its Underground
Economy Two Sides to Every Story
0.3719 0.3694 0.3694 0.3635 Big Data Science with the BD2K-LINCS
Data Coordination and Integration Center
0.3119 0.3131 0.3021 0.2976 Experimental Methods in Systems Biology 0.3391 0.3461 0.3364 0.3358 Dynamical Modeling Methods for Systems
Biology
0.3577 0.364 0.3499 0.3425 The Brain and Space 0.3111 0.3106 0.3083 0.3096 Network Analysis in Systems Biology 0.3183 0.3216 0.32 0.3132
automata 0.5252 0.5312 0.5197 0.515
Natural Language Processing 2013 0.281 0.2744 0.2787 0.2541 Beginning Game Programming with C 0.3215 0.3286 0.3178 0.3069
Climate Change 0.3254 0.3265 0.3283 0.3214
Journalism Skills for Engaged Citizens 0.3496 0.3482 0.3542 0.3472 Algorithms Design and Analysis Part 1 0.3147 0.3108 0.3152 0.2971
136
Algorithms Design and Analysis Part 2 0.3434 0.3404 0.3426 0.3319 Introduction to Chemistry Reactions and
Ratios
0.3002 0.3037 0.2942 0.2893 Genomic and Precision Medicine 0.2923 0.294 0.2877 0.2741 Epigenetic Control of Gene Expression 0.3693 0.3662 0.3671 0.3605 Take the Lead on Healthcare Quality
Improvement
0.4162 0.409 0.4106 0.3959 Surviving Disruptive Technologies 0.4141 0.4025 0.4145 0.4195 Caries Management by Risk Assessment
CAMBRA
0.4162 0.419 0.4026 0.4219 Computational Neuroscience 0.3211 0.3176 0.322 0.3337 Introduction to Data Science 0.2688 0.2678 0.268 0.2597
Discrete Optimization 0.3 0.3047 0.3007 0.2993
Foundations of Virtual Instruction 0.292 0.2838 0.2877 0.255 Virology I How Viruses Work 0.2461 0.2439 0.24 0.2454 Edx Introduction to Computer Science 0.2993 0.298 0.3003 0.2832 edx Big Data in Education 0.4736 0.4731 0.4659 0.2605 edx Cellular mechanisms of brain function 0.1765 0.1768 0.1765 0.176 edx Introduction to Programming with Java
Part 1 Starting to Code with Java
0.1336 0.1318 0.1313 0.122 edx Programming Basics 0.3714 0.3729 0.3708 0.1559
總平均 0.3303 0.3299 0.3273 0.3102
平均 F1 值:與線上摘要器 SweSum 呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比線上摘要器 SweSum 高
在表 57 的結果中除了 Web Application Architectures 課程、Audio Signal Processing for Music Applications 課程、Surviving Disruptive Technologies 課程、
Caries Management by Risk Assessment CAMBRA 課 程 和 Computational Neuroscience 課程的 TF 方法、TF-ISF 方法和 TF-SF 方法平均 F1 值皆比線上摘 要器 SweSum 低,其餘課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法 平均 F1 值至少有一個方法比線上摘要器 SweSum 高。
刪章節使用提示詞 TF 權重正規化加權提示詞時各摘要方法與 PLSA 摘要方 法和一般 LDA 摘要方法比較結果如表 58 所示,與名詞 LDA 摘要方法和動詞 LDA 摘要方法比較結果如表 59 所示,與線上摘要器 SweSum 比較結果如表 60 所示。
表 58. 與 PLSA 摘要方法和一般 LDA 摘要方法比較結果
課程名稱 一般 TF TF-ISF TF-SF PLSA
137 LDA
The Hardware/Software Interface 0.3123 0.3226 0.328 0.3106 0.3151 Social Network Analysis 0.3543 0.3446 0.355 0.335 0.3322 Web Application Architectures 0.4432 0.4363 0.4386 0.4292 0.4318 Audio Signal Processing for Music
Applications
0.1734 0.1926 0.1965 0.1853 0.1791 Malicious Software and its
Underground Economy Two Sides to Every Story
0.3539 0.3492 0.3592 0.3315 0.3361
Big Data Science with the BD2K-LINCS Data Coordination and Integration Center
0.2981 0.2964 0.3046 0.2867 0.2849
Experimental Methods in Systems Biology
0.3241 0.3338 0.3403 0.3223 0.3332 Dynamical Modeling Methods for
Systems Biology
0.3412 0.3476 0.3564 0.3325 0.3309 The Brain and Space 0.2833 0.2978 0.3119 0.2829 0.2978 Network Analysis in Systems
Biology
0.3145 0.3161 0.3191 0.3088 0.3151
automata 0.5121 0.5302 0.5307 0.5274 0.5083
Natural Language Processing 2013 0.2678 0.2763 0.2737 0.2671 0.2664 Beginning Game Programming
with C
0.2788 0.3121 0.3208 0.2927 0.2765 Climate Change 0.2987 0.3253 0.3222 0.3061 0.2851 Journalism Skills for Engaged
Citizens
0.3491 0.3329 0.3357 0.3262 0.3264 Algorithms Design and Analysis
Part 1
0.284 0.307 0.3114 0.2968 0.2876 Algorithms Design and Analysis
Part 2
0.3198 0.3395 0.3407 0.3354 0.3265 Introduction to Chemistry
Reactions and Ratios
0.2841 0.291 0.3024 0.2676 0.275 Genomic and Precision Medicine 0.2793 0.287 0.2936 0.2823 0.2672 Epigenetic Control of Gene Surviving Disruptive Technologies 0.4028 0.3891 0.3978 0.3714 0.3892 Caries Management by Risk
Assessment CAMBRA
0.3657 0.3812 0.4052 0.3504 0.3429 Computational Neuroscience 0.3143 0.3161 0.3205 0.3048 0.3084 Introduction to Data Science 0.2511 0.2582 0.2649 0.2475 0.2563 Discrete Optimization 0.2928 0.2954 0.301 0.2781 0.2846 Foundations of Virtual Instruction 0.2769 0.2469 0.2781 0.221 0.232 Virology I How Viruses Work 0.2343 0.2416 0.2431 0.2279 0.2282 Edx Introduction to Computer
Science edx Introduction to Programming
with Java Part 1 Starting to Code
0.1262 0.1299 0.1294 0.1253 0.1324
138 with Java
edx Programming Basics 0.3772 0.3693 0.373 0.3675 0.3794 總平均 0.3145 0.3198 0.326 0.3074 0.3078
平均 F1 值:與一般 LDA 摘要方法呈現顯著差異
平均 F1 值:與 PLSA 摘要方法呈現顯著差異
紅色字體平均 F1 值:平均 F1 值未比一般 LDA 摘要方法或 PLSA 摘要方法高
在表 58 的結果中 Web Application Architectures 課程、Journalism Skills for Engaged Citizens 課程、Surviving Disruptive Technologies 課程、edx Cellular mechanisms of brain function 課程、edx Introduction to Programming with Java Part 1 Starting to Code with Java 課程和 edx Programming Basics 課程的 TF 摘要方法、
TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值皆比 PLSA 摘要方法或一般 LDA 摘要方法低,其餘課程的 TF 摘要方法、TF-ISF 摘要方法和 TF-SF 摘要方法平均 F1 值至少有一個摘要方法比 PLSA 摘要方法和一般 LDA 摘要方法高。
表 59. 與名詞 LDA 摘要方法和動詞 LDA 摘要方法比較結果
課程名稱 動詞
LDA
TF TF-ISF TF-SF 名詞 LDA The Hardware/Software Interface 0.316 0.3226 0.328 0.3106 0.3171 Social Network Analysis 0.3465 0.3446 0.355 0.335 0.3461 Web Application Architectures 0.4334 0.4363 0.4386 0.4292 0.4337 Audio Signal Processing for Music
Applications
0.1879 0.1926 0.1965 0.1853 0.1734 Malicious Software and its
Underground Economy Two Sides to Every Story
0.3488 0.3492 0.3592 0.3315 0.3508
Big Data Science with the BD2K-LINCS Data Coordination and Integration Center
0.2933 0.2964 0.3046 0.2867 0.2988
Experimental Methods in Systems Biology
0.3268 0.3338 0.3403 0.3223 0.3314 Dynamical Modeling Methods for
Systems Biology
0.3347 0.3476 0.3564 0.3325 0.3479 The Brain and Space 0.2957 0.2978 0.3119 0.2829 0.2933 Network Analysis in Systems
Biology
0.312 0.3161 0.3191 0.3088 0.3174
automata 0.5086 0.5302 0.5307 0.5274 0.5134
Natural Language Processing 2013 0.2694 0.2763 0.2737 0.2671 0 Beginning Game Programming with
C
0.2924 0.3121 0.3208 0.2927 0.2988 Climate Change 0.312 0.3253 0.3222 0.3061 0.3029
139 Journalism Skills for Engaged
Citizens
0.3411 0.3329 0.3357 0.3262 0.3332 Algorithms Design and Analysis Part
1
0.2863 0.307 0.3114 0.2968 0.2879 Algorithms Design and Analysis Part
2
0.3201 0.3395 0.3407 0.3354 0.3166 Introduction to Chemistry Reactions
and Ratios
0.2773 0.291 0.3024 0.2676 0.2783 Genomic and Precision Medicine 0.2588 0.287 0.2936 0.2823 0.2637 Epigenetic Control of Gene
Expression
0.3367 0.3583 0.3618 0.3353 0.335 Take the Lead on Healthcare Quality
Improvement
0.3962 0.3973 0.4041 0.3836 0.3959 Surviving Disruptive Technologies 0.4028 0.3891 0.3978 0.3714 0.407 Caries Management by Risk
Assessment CAMBRA
0.3761 0.3812 0.4052 0.3504 0.3604 Computational Neuroscience 0.3194 0.3161 0.3205 0.3048 0.3044 Introduction to Data Science 0.2506 0.2582 0.2649 0.2475 0.2504 Discrete Optimization 0.2895 0.2954 0.301 0.2781 0.2938 Foundations of Virtual Instruction 0.2665 0.2469 0.2781 0.221 0.2478 Virology I How Viruses Work 0.2343 0.2416 0.2431 0.2279 0.2208 Edx Introduction to Computer
Science
0.2836 0.291 0.2953 0.2831 0.287 edx Big Data in Education 0.453 0.4636 0.4673 0.4465 0.4616 edx Cellular mechanisms of brain
function
0.1805 0.1756 0.1768 0.1768 0.1786 edx Introduction to Programming
with Java Part 1 Starting to Code with Java
with Java Part 1 Starting to Code with Java