Algorithm 2. Transportation Events Alignment and Fusion

Input:${\mathit{T}}_{\mathit{w}}$,${\mathit{T}}_{\mathit{n}}$,$\mathit{W}{\mathit{E}}_{\mathit{w}\mathit{e}\mathit{i}\mathit{b}\mathit{o}}$, $\mathit{W}{\mathit{E}}_{\mathit{n}\mathit{e}\mathit{w}\mathit{s}}$

Output: AlignedTopicsList, FusionEventMatrix

//Step 1: aligning the topic clusters in${\mathit{T}}_{\mathit{w}}$and${\mathit{T}}_{\mathit{n}}$, return AlignedTopicsList

For each ${\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)}$ in ${\mathit{T}}_{\mathit{w}}$

WeiboIndex = i

For each ${\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}$ in ${\mathit{T}}_{\mathit{n}}$

If ${\mu }_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)<\mathrm{minTopicDis}$

$\mathrm{minTopicDis}$$= {\mu }_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$

NewsAlignedIndex = j

End for // Find the closest topic cluster

Append (WeiboIndex, NewsAlignedIndex) to AlignedTopicsList

End for // couple closet the topic clusters into pairs

//Step 2: fusing the words in the cross-paired topics, return FusionEventMatrix

For each index pair in AlignedTopicsList

i = WeiboIndex;

j = NewsAlignedIndex

LBn = ${\mu }_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$$- {\mathsf{\sigma }}_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(i\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$

UBw = ${\mu }_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$ + ${\mathsf{\sigma }}_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$ // calculate the low boundary and up boundary for news words

LBw = ${\mu }_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$ − $3{\mathsf{\sigma }}_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$

UBw = ${\mu }_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$ + $3{\mathsf{\sigma }}_R^{*}\left({\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)},{\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}\right)$ // calculate the low boundary and up boundary for weibo words

For each $\mathit{W}{\mathit{E}}_{\mathit{w}}^{\left(\mathit{m},\mathit{i}\right)}$ in ${\mathit{T}}_{\mathit{w}}^{\left(\mathit{i}\right)}$

For each $\mathit{W}{\mathit{E}}_{\mathit{n}}^{\left(\mathit{n},\mathit{j}\right)}$ in ${\mathit{T}}_{\mathit{n}}^{\left(\mathit{j}\right)}$

If $di{s}_R\left(\mathit{W}{\mathit{E}}_{\mathit{w}}^{\left(\mathit{m},\mathit{i}\right)},\mathit{W}{\mathit{E}}_{\mathit{n}}^{\left(\mathit{n},\mathit{j}\right)}\right) < \mathrm{minWordDis}$

minWordDis = $di{s}_R\left(\mathit{W}{\mathit{E}}_{\mathit{w}}^{\left(\mathit{m},\mathit{i}\right)},\mathit{W}{\mathit{E}}_{\mathit{n}}^{\left(\mathit{n},\mathit{j}\right)}\right)$

minWordIndex = n

End for // Find the shortest word embedding distance in coupled clusters

If LBw < minWordDis < UBw // remove anomaly word in Weibo cluster

If LBn < minWordDis < UBn

Append $\mathit{W}{\mathit{E}}_{\mathit{n}}^{\left(\mathit{m}\mathit{i}\mathit{n}\mathit{W}\mathit{o}\mathit{r}\mathit{d}\mathit{I}\mathit{n}\mathit{d}\mathit{e}\mathit{x},\mathit{j}\right)}$ to FusionEventList // append words in News cluster

Else

Append $\mathit{W}{\mathit{E}}_{\mathit{w}}^{\left(\mathit{m},\mathit{i}\right)}$ to FusionEventList // append words in Weibo cluster

End for

Append FusionEventList to FusionEventMatrix

End for