International Core Journal of Engineering 2020-26 | Page 106

in 2018
B. Research front topic identification
Knime, a large-scale data mining platform, was used to
preprocess data of 9467 fund projects including removing
duplicate data and filtering stop auxiliary words and so on.
Using the log likelihood calculation function of the topic
models toolkit under R language, the corresponding curves of
each sub-time window theme number and logarithmic
likelihood value are obtained, as shown in Fig. 1. to 6.
T ABLE Ⅱ. 17 RESEARCH FRONT TOPICS OBTAINED ACCORDING TO FRONT
IDENTIFICATION INDEXES .
Year
2013
2014
2014
2015
2015
2015
2016
2016
2017
2017
2017
2017
2017
2018
2018
2018
2018
The optimal number of topics for the six sub-periods from
2013 to 2018 years is 37, 43, 46, 38, 55 and 39 respectively,
in combination with manual interpretation. PLDA model in
Knime experimental analysis tool was used to calculate the
topic number of six sub-periods (including article-topic
correspondence,
topic-topic
word-word
frequency
relationship), and the centrality and funding time and funding
amount of fund projects related to the every topic were
calculated. Because the result calculated of the six sub-
periods topic number (article-topic correspondence, topic-
topic word-word frequency relationship) were too much to be
described in detail.
The following lists 17 research front topics with
corresponding index values, which were identified according
to the threshold value of front topic identification indexes, as
shown in Table II. Of the 17 research front topics, one in 2013,
two in 2014, three in 2015, two in 2016, five in 2017 and four
84
Total
Topic number Funding amount
Funding
Centrality
number of fund index(US$/project) time(year/project) index
projects
topic_32
7
1904374. 57
6. 43
11
topic_8
15
1618887. 47
5. 8
12
topic_4
25
1204437. 72
5. 68
16
topic_45
14
661743. 29
4. 79
18
topic_4
69
870938. 25
4. 83
22
topic_26
36
923800. 17
5. 19
20
topic_6
71
855676. 96
4. 32
13
topic_37
27
627436. 74
4. 48
9
topic_53
35
729343. 40
4. 06
14
topic_47
35
823584. 06
4. 06
16
topic_38
41
979762. 00
4. 56
19
topic_13
18
646060. 33
4. 33
17
topic_11
40
559347. 83
4. 15
16
topic_30
10
394689. 60
6
7
topic_2
8
4372649. 63
4. 87
12
topic_18
9
353545. 00
5. 11
9
topic_1
23
548895. 61
4. 7
7