R/017A_Clean1.2_Acq_Impact_Alliance_Function.R
In RayKYang/Research_017A_Rewire: Strategy Research
# 1.1
impact_or_not_MnA <- function(row.in.MnA){
# row.in.MnA = 109
acquirer <- MnA$acquirer.cusip[row.in.MnA]
target <- MnA$target.cusip[row.in.MnA]
acq.Date <- as.Date(MnA$date_ann[row.in.MnA])
which_impact <- pair_year[which((pair_year$X1 == target | pair_year$X2 == target)
& pair_year$Ali_Ann_Date < acq.Date
& pair_year$Ali_Expiration_Date > acq.Date),]
result <- ifelse(nrow(which_impact) > 0, TRUE, FALSE)
return(result)
}
# 1.2
update.alliance.pair <- function(row_num_){
# test
# row_num_ <- 127
(impacting.MnA <- impacting.MnAs[row_num_, ])
# info. for search
(acquirer <- impacting.MnA$acquirer.cusip)
(target <- impacting.MnA$target.cusip)
(acq.Date <- as.Date(impacting.MnA$date_ann))
# find the alliance row under impact
target_related_ali <- pair_year[which((pair_year$X1 == target | pair_year$X2 == target)
& pair_year$Ali_Ann_Date < acq.Date
& pair_year$Ali_Expiration_Date > acq.Date), ]
# make pre_MnA and post_MnA alliance pairs
pre_MnA <- target_related_ali
pre_MnA$Ali_Expiration_Date <- acq.Date - 1
post_MnA <- target_related_ali
post_MnA$Ali_Ann_Date <- acq.Date
if(nrow(post_MnA) > 1){
for (i in 1:nrow(post_MnA))
if(post_MnA[i,]$X1 == target){
post_MnA[i,]$X1_UP <- impacting.MnA$acquirer.cusip_UP
} else {
post_MnA[i,]$X2_UP <- impacting.MnA$acquirer.cusip_UP
}
}else{
if(post_MnA$X1 == target){
post_MnA$X1_UP <- impacting.MnA$acquirer.cusip_UP
} else {
post_MnA$X2_UP <- impacting.MnA$acquirer.cusip_UP
}
}
return(rbind(pre_MnA, post_MnA))
}
# 1.3
affected_Alliance_row_by_each_MnA <- function(row.in.MnA){
# row.in.MnA = 200
target <- MnA$target.cusip[row.in.MnA]
acq.Date <- as.Date(MnA$date_ann[row.in.MnA])
row.in.Alliance <- which((pair_year$X1 == target | pair_year$X2 == target)
& pair_year$Ali_Ann_Date < acq.Date
& pair_year$Ali_Expiration_Date > acq.Date)
return(row.in.Alliance)
}
# 2
add_centrality_change <- function(row_){
# row_ <- 166 # test
df <- impacting.MnAs[row_, ]
acquir_UP <- df$acquirer.cusip_UP
target_UP <- df$target.cusip_UP
pre_event.date <- as.Date(df$date_ann - 1)
event.date <- as.Date(df$date_ann)
# for test only
# pair_year_updated[which(pair_year_updated$Ali_Ann_Date == event.date),]
# pair_year_updated[which(pair_year_updated$Ali_Expiration_Date == pre_event.date),]
# pair_test <- rbind(pair_year_updated[which(pair_year_updated$Ali_Ann_Date == event.date),], pair_year_updated[which(pair_year_updated$Ali_Expiration_Date == pre_event.date),])
# pair_test[which(pair_test$Ali_Ann_Date <= pre_event.date & pair_test$Ali_Expiration_Date >= pre_event.date), ]
# pair_test[which(pair_test$Ali_Ann_Date <= event.date & pair_test$Ali_Expiration_Date >= event.date ), ]
Impact_MnA_on_Firms <- function(a_pre_network, a_post_network, prefix){
centrality <- function(test){
# test <- pre_Ali_Pairs # test <- pos_Ali_Pairs
a <- test[, c("X1_UP", "X2_UP")]
vec <- c(t(a))
g <- igraph::make_graph(vec, directed = F) %>% igraph::simplify()
adjacency_matrix <- as.matrix(igraph::as_adjacency_matrix(g))
Acquirer_from_Outside <- sum(names(igraph::V(g)) == acquir_UP) == 0
Target_from_Outside <- sum(names(igraph::V(g)) == target_UP) == 0
acquir_position <- if(Acquirer_from_Outside){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[acquir_UP,]}
target_position <- if(Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[target_UP,]}
result <- data.frame(cusipAup = names(igraph::V(g)),
# Add notes
Acquirer_from_Outside = ifelse(Acquirer_from_Outside == TRUE, "Y", "N"),
Target_from_Outside = ifelse(Target_from_Outside == TRUE, "Y", "N"),
# ego network cohesion
ego_nw_cohesion = unlist(purrr::map(igraph::make_ego_graph(g, order = 1, nodes = igraph::V(g)), igraph::cohesion)),
# centrality
degree=igraph::degree(g),
eigen=igraph::eigen_centrality(g)$vector,
btwness=igraph::betweenness(g),
constraint=igraph::constraint(g),
# shortest path (geodesic)
shortest.path.to.acquir = if(Acquirer_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
as.vector(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == acquir_UP)))},
shortest.path.to.target = if(Target_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
as.vector(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == target_UP)))},
shortest.path_acquir_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
rep(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == target_UP))[which(names(igraph::V(g)) == acquir_UP)], length(names(igraph::V(g))))},
# positional similarity (correlation)
positional_correlation_to_acquir = if(Acquirer_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[acquir_UP,]},
positional_correlation_to_target = if(Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[target_UP,]},
positional_correlation_Acquirer_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[target_UP,][acquir_UP]},
# positional similarity (euclidean)
positional_euclidean_to_acquir = if(Acquirer_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - acquir_position)^2)))},
positional_euclidean_to_target = if(Target_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - target_position)^2)))},
positional_euclidean_Acquirer_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - target_position)^2)))[acquir_UP]},
stringsAsFactors = FALSE)
return(result)
}
# test
# a_pre_network <- pre_Ali_Pairs
# a_post_network <- pos_Ali_Pairs
# prefix <- "GENERAL_"
pre_centrality <- centrality(a_pre_network)
pos_centrality <- centrality(a_post_network)
# replace infinite with max value
pre_centrality$shortest.path.to.acquir[is.infinite(pre_centrality$shortest.path.to.acquir)] <- NA
pos_centrality$shortest.path.to.acquir[is.infinite(pos_centrality$shortest.path.to.acquir)] <- NA
pre_centrality$shortest.path.to.target[is.infinite(pre_centrality$shortest.path.to.target)] <- NA
pos_centrality$shortest.path.to.target[is.infinite(pos_centrality$shortest.path.to.target)] <- NA
pre_centrality$shortest.path_acquir_to_target[is.infinite(pre_centrality$shortest.path_acquir_to_target)] <- NA
pos_centrality$shortest.path_acquir_to_target[is.infinite(pos_centrality$shortest.path_acquir_to_target)] <- NA
max.value <- max(max(pre_centrality$shortest.path.to.acquir, na.rm = TRUE),
max(pre_centrality$shortest.path.to.target, na.rm = TRUE),
max(pos_centrality$shortest.path.to.acquir, na.rm = TRUE),
max(pos_centrality$shortest.path.to.target, na.rm = TRUE))
pre_centrality$shortest.path.to.acquir[is.na(pre_centrality$shortest.path.to.acquir)] <- max.value + 1
pos_centrality$shortest.path.to.acquir[is.na(pos_centrality$shortest.path.to.acquir)] <- max.value + 1
pre_centrality$shortest.path.to.target[is.na(pre_centrality$shortest.path.to.target)] <- max.value + 1
pos_centrality$shortest.path.to.target[is.na(pos_centrality$shortest.path.to.target)] <- max.value + 1
pre_centrality$shortest.path_acquir_to_target[is.na(pre_centrality$shortest.path_acquir_to_target)] <- max.value + 1
pos_centrality$shortest.path_acquir_to_target[is.na(pos_centrality$shortest.path_acquir_to_target)] <- max.value + 1
max.value_euc <- max(max(pre_centrality$positional_euclidean_to_acquir, na.rm = TRUE),
max(pre_centrality$positional_euclidean_to_target, na.rm = TRUE),
max(pos_centrality$positional_euclidean_to_acquir, na.rm = TRUE),
max(pos_centrality$positional_euclidean_to_target, na.rm = TRUE))
pre_centrality$positional_euclidean_to_acquir[is.na(pre_centrality$positional_euclidean_to_acquir)] <- max.value_euc + 1
pre_centrality$positional_euclidean_to_target[is.na(pre_centrality$positional_euclidean_to_target)] <- max.value_euc + 1
pre_centrality$positional_euclidean_Acquirer_to_target[is.na(pre_centrality$positional_euclidean_Acquirer_to_target)] <- max.value_euc + 1
pos_centrality$positional_euclidean_to_acquir[is.na(pos_centrality$positional_euclidean_to_acquir)] <- max.value_euc + 1
pos_centrality$positional_euclidean_to_target[is.na(pos_centrality$positional_euclidean_to_target)] <- max.value_euc + 1
pos_centrality$positional_euclidean_Acquirer_to_target[is.na(pos_centrality$positional_euclidean_Acquirer_to_target)] <- max.value_euc + 1
names(pre_centrality)[2:17] <- paste0("pre_", names(pre_centrality)[2:17])
# delta centrality #
merged_ <- merge(pre_centrality, pos_centrality, by = "cusipAup")
merged_$pre_btwness <- regrrr::scale_01(merged_$pre_btwness)
merged_$btwness <- regrrr::scale_01(merged_$btwness)
merged_$delta_degree <- merged_$degree - merged_$pre_degree
merged_$delta_eigen <- merged_$eigen - merged_$pre_eigen
merged_$delta_btwness <- merged_$btwness - merged_$pre_btwness
merged_$delta_strhole <- merged_$pre_constraint - merged_$constraint
merged_$delta_nw_cohesion <- merged_$ego_nw_cohesion - merged_$pre_ego_nw_cohesion
merged_$delta_spath2acquir <- merged_$shortest.path.to.acquir - merged_$pre_shortest.path.to.acquir
merged_$delta_spath2target <- merged_$shortest.path.to.target - merged_$pre_shortest.path.to.target
# delta position #
merged_$delta_cor_to_acquir <- merged_$positional_correlation_to_acquir - merged_$pre_positional_correlation_to_acquir
merged_$delta_cor_to_target <- merged_$positional_correlation_to_target - merged_$pre_positional_correlation_to_target
merged_$delta_euc_to_acquir <- merged_$positional_euclidean_to_acquir - merged_$pre_positional_euclidean_to_acquir
merged_$delta_euc_to_target <- merged_$positional_euclidean_to_target - merged_$pre_positional_euclidean_to_target
# merged_ <- merged_ %>% dplyr::filter(abs(delta_eigen) > 0 | abs(delta_strhole) > 0)
names(merged_)[2:length(names(merged_))] <- paste0(prefix, names(merged_)[2:length(names(merged_))])
return(merged_)}
# 1. general alliance network #
pre_Ali_Pairs <- pair_year_updated[which(pair_year_updated$Ali_Ann_Date <= pre_event.date & pair_year_updated$Ali_Expiration_Date >= pre_event.date), ]
pos_Ali_Pairs <- pair_year_updated[which(pair_year_updated$Ali_Ann_Date <= event.date & pair_year_updated$Ali_Expiration_Date >= event.date ), ]
impact_general <- Impact_MnA_on_Firms(pre_Ali_Pairs, pos_Ali_Pairs, "GENERAL_")
# dim(impact_general)
# # 2.a JV network #
Impact_MnA_on_SIMPLE <- function(a_pre_network, a_post_network, prefix){
centrality <- function(test){
# test <- pre_Ali_Pairs # test <- pos_Ali_Pairs
a <- test[, c("X1_UP", "X2_UP")]
vec <- c(t(a))
g <- igraph::make_graph(vec, directed = F) %>% igraph::simplify()
adjacency_matrix <- as.matrix(igraph::as_adjacency_matrix(g))
acquir_position <- if(sum(names(igraph::V(g)) == acquir_UP) == 0){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[acquir_UP,]}
target_position <- if(sum(names(igraph::V(g)) == target_UP) == 0){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[target_UP,]}
Acquirer_from_Outside <- sum(names(igraph::V(g)) == acquir_UP) == 0
Target_from_Outside <- sum(names(igraph::V(g)) == target_UP) == 0
result <- data.frame(cusipAup = names(igraph::V(g)),
# Add notes
Acquirer_from_Outside = ifelse(Acquirer_from_Outside == TRUE, "Y", "N"),
Target_from_Outside = ifelse(Target_from_Outside == TRUE, "Y", "N"),
# ego network cohesion
ego_nw_cohesion = unlist(purrr::map(igraph::make_ego_graph(g, order = 1, nodes = igraph::V(g)), igraph::cohesion)),
# centrality
degree=igraph::degree(g),
eigen=igraph::eigen_centrality(g)$vector,
btwness=igraph::betweenness(g),
constraint=igraph::constraint(g),
stringsAsFactors = FALSE)
return(result)
}
# test
# a_pre_network <- pre_Ali_Pairs
# a_post_network <- pos_Ali_Pairs
# prefix <- "GENERAL_"
pre_centrality <- centrality(a_pre_network)
pos_centrality <- centrality(a_post_network)
names(pre_centrality)[2:length(names(pre_centrality))] <- paste0("pre_", names(pre_centrality)[2:length(names(pre_centrality))])
# delta centrality #
merged_ <- merge(pre_centrality, pos_centrality, by = "cusipAup")
merged_$pre_btwness <- regrrr::scale_01(merged_$pre_btwness)
merged_$btwness <- regrrr::scale_01(merged_$btwness)
merged_$delta_degree <- merged_$degree - merged_$pre_degree
merged_$delta_eigen <- merged_$eigen - merged_$pre_eigen
merged_$delta_btwness <- merged_$btwness - merged_$pre_btwness
merged_$delta_strhole <- merged_$pre_constraint - merged_$constraint
merged_$delta_nw_cohesion <- merged_$ego_nw_cohesion - merged_$pre_ego_nw_cohesion
names(merged_)[2:length(names(merged_))] <- paste0(prefix, names(merged_)[2:length(names(merged_))])
return(merged_)}
pre_JV_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$JV_Flag == "Y"), ]
pos_JV_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$JV_Flag == "Y"), ]
if(nrow(pre_JV_Pairs) > 0 && nrow(pos_JV_Pairs) > 0){
impact_JV <- Impact_MnA_on_Firms(pre_JV_Pairs, pos_JV_Pairs, "JV_")
impact_general <- dplyr::left_join(impact_general, impact_JV, by = "cusipAup")
}
# dim(impact_general)
# # 2.b RnD network #
pre_RnD_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$RnD_Flag == "Y"), ]
pos_RnD_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$RnD_Flag == "Y"), ]
if(nrow(pre_RnD_Pairs) > 0 && nrow(pos_RnD_Pairs) > 0){
impact_RnD <- Impact_MnA_on_Firms(pre_RnD_Pairs, pos_RnD_Pairs, "RnD_")
impact_general <- dplyr::left_join(impact_general, impact_RnD, by = "cusipAup")
}
# # 2.c Technology network #
pre_Technology_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$Technology_Transfer == "Y"), ]
pos_Technology_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$Technology_Transfer == "Y"), ]
if(nrow(pre_Technology_Pairs) > 0 && nrow(pos_Technology_Pairs) > 0){
impact_Technology <- Impact_MnA_on_Firms(pre_Technology_Pairs, pos_Technology_Pairs, "Technology_")
impact_general <- dplyr::left_join(impact_general, impact_Technology, by = "cusipAup")
}
##
if(nrow(impact_general) > 0){impact_general$event_number <- df$event_number}
print(row_)
print(dim(impact_general))
print(df)
return(impact_general)}
# 2.2
add_centrality_change_2.2 <- function(row_){
# row_ <- 1 # test
df <- MnA[row_, ]
acquir_UP <- df$acquirer.cusip_UP
target_UP <- df$target.cusip_UP
pre_event.date <- as.Date(df$date_ann - 1)
event.date <- as.Date(df$date_ann)
# for test only
# pair_year_updated[which(pair_year_updated$Ali_Ann_Date == event.date),]
# pair_year_updated[which(pair_year_updated$Ali_Expiration_Date == pre_event.date),]
# pair_test <- rbind(pair_year_updated[which(pair_year_updated$Ali_Ann_Date == event.date),], pair_year_updated[which(pair_year_updated$Ali_Expiration_Date == pre_event.date),])
# pair_test[which(pair_test$Ali_Ann_Date <= pre_event.date & pair_test$Ali_Expiration_Date >= pre_event.date), ]
# pair_test[which(pair_test$Ali_Ann_Date <= event.date & pair_test$Ali_Expiration_Date >= event.date ), ]
Impact_MnA_on_Firms <- function(a_pre_network, a_post_network, prefix){
centrality <- function(test){
# test <- pre_Ali_Pairs # test <- pos_Ali_Pairs
a <- test[, c("X1_UP", "X2_UP")]
vec <- c(t(a))
g <- igraph::make_graph(vec, directed = F) %>% igraph::simplify()
adjacency_matrix <- as.matrix(igraph::as_adjacency_matrix(g))
Acquirer_from_Outside <- sum(names(igraph::V(g)) == acquir_UP) == 0
Target_from_Outside <- sum(names(igraph::V(g)) == target_UP) == 0
acquir_position <- if(Acquirer_from_Outside){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[acquir_UP,]}
target_position <- if(Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[target_UP,]}
result <- data.frame(cusipAup = names(igraph::V(g)),
# Add notes
Acquirer_from_Outside = ifelse(Acquirer_from_Outside == TRUE, "Y", "N"),
Target_from_Outside = ifelse(Target_from_Outside == TRUE, "Y", "N"),
# ego network cohesion
ego_nw_cohesion = unlist(purrr::map(igraph::make_ego_graph(g, order = 1, nodes = igraph::V(g)), igraph::cohesion)),
# centrality
degree=igraph::degree(g),
eigen=igraph::eigen_centrality(g)$vector,
btwness=igraph::betweenness(g),
constraint=igraph::constraint(g),
# shortest path (geodesic)
shortest.path.to.acquir = if(Acquirer_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
as.vector(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == acquir_UP)))},
shortest.path.to.target = if(Target_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
as.vector(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == target_UP)))},
shortest.path_acquir_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
rep(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == target_UP))[which(names(igraph::V(g)) == acquir_UP)], length(names(igraph::V(g))))},
# positional similarity (correlation)
positional_correlation_to_acquir = if(Acquirer_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[acquir_UP,]},
positional_correlation_to_target = if(Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[target_UP,]},
positional_correlation_Acquirer_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[target_UP,][acquir_UP]},
# positional similarity (euclidean)
positional_euclidean_to_acquir = if(Acquirer_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - acquir_position)^2)))},
positional_euclidean_to_target = if(Target_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - target_position)^2)))},
positional_euclidean_Acquirer_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - target_position)^2)))[acquir_UP]},
stringsAsFactors = FALSE)
return(result)
}
# test
# a_pre_network <- pre_Ali_Pairs
# a_post_network <- pos_Ali_Pairs
# prefix <- "GENERAL_"
pre_centrality <- centrality(a_pre_network)
pos_centrality <- centrality(a_post_network)
# replace infinite with max value
pre_centrality$shortest.path.to.acquir[is.infinite(pre_centrality$shortest.path.to.acquir)] <- NA
pos_centrality$shortest.path.to.acquir[is.infinite(pos_centrality$shortest.path.to.acquir)] <- NA
pre_centrality$shortest.path.to.target[is.infinite(pre_centrality$shortest.path.to.target)] <- NA
pos_centrality$shortest.path.to.target[is.infinite(pos_centrality$shortest.path.to.target)] <- NA
pre_centrality$shortest.path_acquir_to_target[is.infinite(pre_centrality$shortest.path_acquir_to_target)] <- NA
pos_centrality$shortest.path_acquir_to_target[is.infinite(pos_centrality$shortest.path_acquir_to_target)] <- NA
max.value <- max(max(pre_centrality$shortest.path.to.acquir, na.rm = TRUE),
max(pre_centrality$shortest.path.to.target, na.rm = TRUE),
max(pos_centrality$shortest.path.to.acquir, na.rm = TRUE),
max(pos_centrality$shortest.path.to.target, na.rm = TRUE))
pre_centrality$shortest.path.to.acquir[is.na(pre_centrality$shortest.path.to.acquir)] <- max.value + 1
pos_centrality$shortest.path.to.acquir[is.na(pos_centrality$shortest.path.to.acquir)] <- max.value + 1
pre_centrality$shortest.path.to.target[is.na(pre_centrality$shortest.path.to.target)] <- max.value + 1
pos_centrality$shortest.path.to.target[is.na(pos_centrality$shortest.path.to.target)] <- max.value + 1
pre_centrality$shortest.path_acquir_to_target[is.na(pre_centrality$shortest.path_acquir_to_target)] <- max.value + 1
pos_centrality$shortest.path_acquir_to_target[is.na(pos_centrality$shortest.path_acquir_to_target)] <- max.value + 1
max.value_euc <- max(max(pre_centrality$positional_euclidean_to_acquir, na.rm = TRUE),
max(pre_centrality$positional_euclidean_to_target, na.rm = TRUE),
max(pos_centrality$positional_euclidean_to_acquir, na.rm = TRUE),
max(pos_centrality$positional_euclidean_to_target, na.rm = TRUE))
pre_centrality$positional_euclidean_to_acquir[is.na(pre_centrality$positional_euclidean_to_acquir)] <- max.value_euc + 1
pre_centrality$positional_euclidean_to_target[is.na(pre_centrality$positional_euclidean_to_target)] <- max.value_euc + 1
pre_centrality$positional_euclidean_Acquirer_to_target[is.na(pre_centrality$positional_euclidean_Acquirer_to_target)] <- max.value_euc + 1
pos_centrality$positional_euclidean_to_acquir[is.na(pos_centrality$positional_euclidean_to_acquir)] <- max.value_euc + 1
pos_centrality$positional_euclidean_to_target[is.na(pos_centrality$positional_euclidean_to_target)] <- max.value_euc + 1
pos_centrality$positional_euclidean_Acquirer_to_target[is.na(pos_centrality$positional_euclidean_Acquirer_to_target)] <- max.value_euc + 1
names(pre_centrality)[2:17] <- paste0("pre_", names(pre_centrality)[2:17])
# delta centrality #
merged_ <- merge(pre_centrality, pos_centrality, by = "cusipAup")
merged_$pre_btwness <- regrrr::scale_01(merged_$pre_btwness)
merged_$btwness <- regrrr::scale_01(merged_$btwness)
merged_$delta_degree <- merged_$degree - merged_$pre_degree
merged_$delta_eigen <- merged_$eigen - merged_$pre_eigen
merged_$delta_btwness <- merged_$btwness - merged_$pre_btwness
merged_$delta_strhole <- merged_$pre_constraint - merged_$constraint
merged_$delta_nw_cohesion <- merged_$ego_nw_cohesion - merged_$pre_ego_nw_cohesion
merged_$delta_spath2acquir <- merged_$shortest.path.to.acquir - merged_$pre_shortest.path.to.acquir
merged_$delta_spath2target <- merged_$shortest.path.to.target - merged_$pre_shortest.path.to.target
# delta position #
merged_$delta_cor_to_acquir <- merged_$positional_correlation_to_acquir - merged_$pre_positional_correlation_to_acquir
merged_$delta_cor_to_target <- merged_$positional_correlation_to_target - merged_$pre_positional_correlation_to_target
merged_$delta_euc_to_acquir <- merged_$positional_euclidean_to_acquir - merged_$pre_positional_euclidean_to_acquir
merged_$delta_euc_to_target <- merged_$positional_euclidean_to_target - merged_$pre_positional_euclidean_to_target
merged_ <- merged_ %>% dplyr::filter(abs(delta_eigen) > 0.001 | abs(delta_strhole) > 0.001)
names(merged_)[2:length(names(merged_))] <- paste0(prefix, names(merged_)[2:length(names(merged_))])
return(merged_)}
# 1. general alliance network #
pre_Ali_Pairs <- pair_year_updated[which(pair_year_updated$Ali_Ann_Date <= pre_event.date & pair_year_updated$Ali_Expiration_Date >= pre_event.date), ]
pos_Ali_Pairs <- pair_year_updated[which(pair_year_updated$Ali_Ann_Date <= event.date & pair_year_updated$Ali_Expiration_Date >= event.date ), ]
if(nrow(pre_Ali_Pairs) > 0 && nrow(pos_Ali_Pairs) > 0){
impact_general <- Impact_MnA_on_Firms(pre_Ali_Pairs, pos_Ali_Pairs, "GENERAL_")
}
# dim(impact_general)
# # 2.a JV network #
pre_JV_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$JV_Flag == "Y"), ]
pos_JV_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$JV_Flag == "Y"), ]
if(nrow(pre_JV_Pairs) > 0 && nrow(pos_JV_Pairs) > 0){
impact_JV <- Impact_MnA_on_Firms(pre_JV_Pairs, pos_JV_Pairs, "JV_")
impact_general <- dplyr::left_join(impact_general, impact_JV, by = "cusipAup")
}
# dim(impact_general)
# # 2.b RnD network #
pre_RnD_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$RnD_Flag == "Y"), ]
pos_RnD_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$RnD_Flag == "Y"), ]
if(nrow(pre_RnD_Pairs) > 0 && nrow(pos_RnD_Pairs) > 0){
impact_RnD <- Impact_MnA_on_Firms(pre_RnD_Pairs, pos_RnD_Pairs, "RnD_")
impact_general <- dplyr::left_join(impact_general, impact_RnD, by = "cusipAup")
}
# # 2.c Technology network #
pre_Technology_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$Technology_Transfer == "Y"), ]
pos_Technology_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$Technology_Transfer == "Y"), ]
if(nrow(pre_Technology_Pairs) > 0 && nrow(pos_Technology_Pairs) > 0){
impact_Technology <- Impact_MnA_on_Firms(pre_Technology_Pairs, pos_Technology_Pairs, "Technology_")
impact_general <- dplyr::left_join(impact_general, impact_Technology, by = "cusipAup")
}
##
if(nrow(pre_Ali_Pairs) > 0){
if(nrow(impact_general) > 0){impact_general$impacting_or_not <- df$impacting_or_not}
}
print(row_)
return(impact_general)}
RayKYang/Research_017A_Rewire documentation built on May 6, 2019, 7:05 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
# 1.1
impact_or_not_MnA <- function(row.in.MnA){
# row.in.MnA = 109
acquirer <- MnA$acquirer.cusip[row.in.MnA]
target <- MnA$target.cusip[row.in.MnA]
acq.Date <- as.Date(MnA$date_ann[row.in.MnA])
which_impact <- pair_year[which((pair_year$X1 == target | pair_year$X2 == target)
& pair_year$Ali_Ann_Date < acq.Date
& pair_year$Ali_Expiration_Date > acq.Date),]
result <- ifelse(nrow(which_impact) > 0, TRUE, FALSE)
return(result)
}
# 1.2
update.alliance.pair <- function(row_num_){
# test
# row_num_ <- 127
(impacting.MnA <- impacting.MnAs[row_num_, ])
# info. for search
(acquirer <- impacting.MnA$acquirer.cusip)
(target <- impacting.MnA$target.cusip)
(acq.Date <- as.Date(impacting.MnA$date_ann))
# find the alliance row under impact
target_related_ali <- pair_year[which((pair_year$X1 == target | pair_year$X2 == target)
& pair_year$Ali_Ann_Date < acq.Date
& pair_year$Ali_Expiration_Date > acq.Date), ]
# make pre_MnA and post_MnA alliance pairs
pre_MnA <- target_related_ali
pre_MnA$Ali_Expiration_Date <- acq.Date - 1
post_MnA <- target_related_ali
post_MnA$Ali_Ann_Date <- acq.Date
if(nrow(post_MnA) > 1){
for (i in 1:nrow(post_MnA))
if(post_MnA[i,]$X1 == target){
post_MnA[i,]$X1_UP <- impacting.MnA$acquirer.cusip_UP
} else {
post_MnA[i,]$X2_UP <- impacting.MnA$acquirer.cusip_UP
}
}else{
if(post_MnA$X1 == target){
post_MnA$X1_UP <- impacting.MnA$acquirer.cusip_UP
} else {
post_MnA$X2_UP <- impacting.MnA$acquirer.cusip_UP
}
}
return(rbind(pre_MnA, post_MnA))
}
# 1.3
affected_Alliance_row_by_each_MnA <- function(row.in.MnA){
# row.in.MnA = 200
target <- MnA$target.cusip[row.in.MnA]
acq.Date <- as.Date(MnA$date_ann[row.in.MnA])
row.in.Alliance <- which((pair_year$X1 == target | pair_year$X2 == target)
& pair_year$Ali_Ann_Date < acq.Date
& pair_year$Ali_Expiration_Date > acq.Date)
return(row.in.Alliance)
}
# 2
add_centrality_change <- function(row_){
# row_ <- 166 # test
df <- impacting.MnAs[row_, ]
acquir_UP <- df$acquirer.cusip_UP
target_UP <- df$target.cusip_UP
pre_event.date <- as.Date(df$date_ann - 1)
event.date <- as.Date(df$date_ann)
# for test only
# pair_year_updated[which(pair_year_updated$Ali_Ann_Date == event.date),]
# pair_year_updated[which(pair_year_updated$Ali_Expiration_Date == pre_event.date),]
# pair_test <- rbind(pair_year_updated[which(pair_year_updated$Ali_Ann_Date == event.date),], pair_year_updated[which(pair_year_updated$Ali_Expiration_Date == pre_event.date),])
# pair_test[which(pair_test$Ali_Ann_Date <= pre_event.date & pair_test$Ali_Expiration_Date >= pre_event.date), ]
# pair_test[which(pair_test$Ali_Ann_Date <= event.date & pair_test$Ali_Expiration_Date >= event.date ), ]
Impact_MnA_on_Firms <- function(a_pre_network, a_post_network, prefix){
centrality <- function(test){
# test <- pre_Ali_Pairs # test <- pos_Ali_Pairs
a <- test[, c("X1_UP", "X2_UP")]
vec <- c(t(a))
g <- igraph::make_graph(vec, directed = F) %>% igraph::simplify()
adjacency_matrix <- as.matrix(igraph::as_adjacency_matrix(g))
Acquirer_from_Outside <- sum(names(igraph::V(g)) == acquir_UP) == 0
Target_from_Outside <- sum(names(igraph::V(g)) == target_UP) == 0
acquir_position <- if(Acquirer_from_Outside){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[acquir_UP,]}
target_position <- if(Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[target_UP,]}
result <- data.frame(cusipAup = names(igraph::V(g)),
# Add notes
Acquirer_from_Outside = ifelse(Acquirer_from_Outside == TRUE, "Y", "N"),
Target_from_Outside = ifelse(Target_from_Outside == TRUE, "Y", "N"),
# ego network cohesion
ego_nw_cohesion = unlist(purrr::map(igraph::make_ego_graph(g, order = 1, nodes = igraph::V(g)), igraph::cohesion)),
# centrality
degree=igraph::degree(g),
eigen=igraph::eigen_centrality(g)$vector,
btwness=igraph::betweenness(g),
constraint=igraph::constraint(g),
# shortest path (geodesic)
shortest.path.to.acquir = if(Acquirer_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
as.vector(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == acquir_UP)))},
shortest.path.to.target = if(Target_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
as.vector(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == target_UP)))},
shortest.path_acquir_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
rep(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == target_UP))[which(names(igraph::V(g)) == acquir_UP)], length(names(igraph::V(g))))},
# positional similarity (correlation)
positional_correlation_to_acquir = if(Acquirer_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[acquir_UP,]},
positional_correlation_to_target = if(Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[target_UP,]},
positional_correlation_Acquirer_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[target_UP,][acquir_UP]},
# positional similarity (euclidean)
positional_euclidean_to_acquir = if(Acquirer_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - acquir_position)^2)))},
positional_euclidean_to_target = if(Target_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - target_position)^2)))},
positional_euclidean_Acquirer_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - target_position)^2)))[acquir_UP]},
stringsAsFactors = FALSE)
return(result)
}
# test
# a_pre_network <- pre_Ali_Pairs
# a_post_network <- pos_Ali_Pairs
# prefix <- "GENERAL_"
pre_centrality <- centrality(a_pre_network)
pos_centrality <- centrality(a_post_network)
# replace infinite with max value
pre_centrality$shortest.path.to.acquir[is.infinite(pre_centrality$shortest.path.to.acquir)] <- NA
pos_centrality$shortest.path.to.acquir[is.infinite(pos_centrality$shortest.path.to.acquir)] <- NA
pre_centrality$shortest.path.to.target[is.infinite(pre_centrality$shortest.path.to.target)] <- NA
pos_centrality$shortest.path.to.target[is.infinite(pos_centrality$shortest.path.to.target)] <- NA
pre_centrality$shortest.path_acquir_to_target[is.infinite(pre_centrality$shortest.path_acquir_to_target)] <- NA
pos_centrality$shortest.path_acquir_to_target[is.infinite(pos_centrality$shortest.path_acquir_to_target)] <- NA
max.value <- max(max(pre_centrality$shortest.path.to.acquir, na.rm = TRUE),
max(pre_centrality$shortest.path.to.target, na.rm = TRUE),
max(pos_centrality$shortest.path.to.acquir, na.rm = TRUE),
max(pos_centrality$shortest.path.to.target, na.rm = TRUE))
pre_centrality$shortest.path.to.acquir[is.na(pre_centrality$shortest.path.to.acquir)] <- max.value + 1
pos_centrality$shortest.path.to.acquir[is.na(pos_centrality$shortest.path.to.acquir)] <- max.value + 1
pre_centrality$shortest.path.to.target[is.na(pre_centrality$shortest.path.to.target)] <- max.value + 1
pos_centrality$shortest.path.to.target[is.na(pos_centrality$shortest.path.to.target)] <- max.value + 1
pre_centrality$shortest.path_acquir_to_target[is.na(pre_centrality$shortest.path_acquir_to_target)] <- max.value + 1
pos_centrality$shortest.path_acquir_to_target[is.na(pos_centrality$shortest.path_acquir_to_target)] <- max.value + 1
max.value_euc <- max(max(pre_centrality$positional_euclidean_to_acquir, na.rm = TRUE),
max(pre_centrality$positional_euclidean_to_target, na.rm = TRUE),
max(pos_centrality$positional_euclidean_to_acquir, na.rm = TRUE),
max(pos_centrality$positional_euclidean_to_target, na.rm = TRUE))
pre_centrality$positional_euclidean_to_acquir[is.na(pre_centrality$positional_euclidean_to_acquir)] <- max.value_euc + 1
pre_centrality$positional_euclidean_to_target[is.na(pre_centrality$positional_euclidean_to_target)] <- max.value_euc + 1
pre_centrality$positional_euclidean_Acquirer_to_target[is.na(pre_centrality$positional_euclidean_Acquirer_to_target)] <- max.value_euc + 1
pos_centrality$positional_euclidean_to_acquir[is.na(pos_centrality$positional_euclidean_to_acquir)] <- max.value_euc + 1
pos_centrality$positional_euclidean_to_target[is.na(pos_centrality$positional_euclidean_to_target)] <- max.value_euc + 1
pos_centrality$positional_euclidean_Acquirer_to_target[is.na(pos_centrality$positional_euclidean_Acquirer_to_target)] <- max.value_euc + 1
names(pre_centrality)[2:17] <- paste0("pre_", names(pre_centrality)[2:17])
# delta centrality #
merged_ <- merge(pre_centrality, pos_centrality, by = "cusipAup")
merged_$pre_btwness <- regrrr::scale_01(merged_$pre_btwness)
merged_$btwness <- regrrr::scale_01(merged_$btwness)
merged_$delta_degree <- merged_$degree - merged_$pre_degree
merged_$delta_eigen <- merged_$eigen - merged_$pre_eigen
merged_$delta_btwness <- merged_$btwness - merged_$pre_btwness
merged_$delta_strhole <- merged_$pre_constraint - merged_$constraint
merged_$delta_nw_cohesion <- merged_$ego_nw_cohesion - merged_$pre_ego_nw_cohesion
merged_$delta_spath2acquir <- merged_$shortest.path.to.acquir - merged_$pre_shortest.path.to.acquir
merged_$delta_spath2target <- merged_$shortest.path.to.target - merged_$pre_shortest.path.to.target
# delta position #
merged_$delta_cor_to_acquir <- merged_$positional_correlation_to_acquir - merged_$pre_positional_correlation_to_acquir
merged_$delta_cor_to_target <- merged_$positional_correlation_to_target - merged_$pre_positional_correlation_to_target
merged_$delta_euc_to_acquir <- merged_$positional_euclidean_to_acquir - merged_$pre_positional_euclidean_to_acquir
merged_$delta_euc_to_target <- merged_$positional_euclidean_to_target - merged_$pre_positional_euclidean_to_target
# merged_ <- merged_ %>% dplyr::filter(abs(delta_eigen) > 0 | abs(delta_strhole) > 0)
names(merged_)[2:length(names(merged_))] <- paste0(prefix, names(merged_)[2:length(names(merged_))])
return(merged_)}
# 1. general alliance network #
pre_Ali_Pairs <- pair_year_updated[which(pair_year_updated$Ali_Ann_Date <= pre_event.date & pair_year_updated$Ali_Expiration_Date >= pre_event.date), ]
pos_Ali_Pairs <- pair_year_updated[which(pair_year_updated$Ali_Ann_Date <= event.date & pair_year_updated$Ali_Expiration_Date >= event.date ), ]
impact_general <- Impact_MnA_on_Firms(pre_Ali_Pairs, pos_Ali_Pairs, "GENERAL_")
# dim(impact_general)
# # 2.a JV network #
Impact_MnA_on_SIMPLE <- function(a_pre_network, a_post_network, prefix){
centrality <- function(test){
# test <- pre_Ali_Pairs # test <- pos_Ali_Pairs
a <- test[, c("X1_UP", "X2_UP")]
vec <- c(t(a))
g <- igraph::make_graph(vec, directed = F) %>% igraph::simplify()
adjacency_matrix <- as.matrix(igraph::as_adjacency_matrix(g))
acquir_position <- if(sum(names(igraph::V(g)) == acquir_UP) == 0){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[acquir_UP,]}
target_position <- if(sum(names(igraph::V(g)) == target_UP) == 0){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[target_UP,]}
Acquirer_from_Outside <- sum(names(igraph::V(g)) == acquir_UP) == 0
Target_from_Outside <- sum(names(igraph::V(g)) == target_UP) == 0
result <- data.frame(cusipAup = names(igraph::V(g)),
# Add notes
Acquirer_from_Outside = ifelse(Acquirer_from_Outside == TRUE, "Y", "N"),
Target_from_Outside = ifelse(Target_from_Outside == TRUE, "Y", "N"),
# ego network cohesion
ego_nw_cohesion = unlist(purrr::map(igraph::make_ego_graph(g, order = 1, nodes = igraph::V(g)), igraph::cohesion)),
# centrality
degree=igraph::degree(g),
eigen=igraph::eigen_centrality(g)$vector,
btwness=igraph::betweenness(g),
constraint=igraph::constraint(g),
stringsAsFactors = FALSE)
return(result)
}
# test
# a_pre_network <- pre_Ali_Pairs
# a_post_network <- pos_Ali_Pairs
# prefix <- "GENERAL_"
pre_centrality <- centrality(a_pre_network)
pos_centrality <- centrality(a_post_network)
names(pre_centrality)[2:length(names(pre_centrality))] <- paste0("pre_", names(pre_centrality)[2:length(names(pre_centrality))])
# delta centrality #
merged_ <- merge(pre_centrality, pos_centrality, by = "cusipAup")
merged_$pre_btwness <- regrrr::scale_01(merged_$pre_btwness)
merged_$btwness <- regrrr::scale_01(merged_$btwness)
merged_$delta_degree <- merged_$degree - merged_$pre_degree
merged_$delta_eigen <- merged_$eigen - merged_$pre_eigen
merged_$delta_btwness <- merged_$btwness - merged_$pre_btwness
merged_$delta_strhole <- merged_$pre_constraint - merged_$constraint
merged_$delta_nw_cohesion <- merged_$ego_nw_cohesion - merged_$pre_ego_nw_cohesion
names(merged_)[2:length(names(merged_))] <- paste0(prefix, names(merged_)[2:length(names(merged_))])
return(merged_)}
pre_JV_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$JV_Flag == "Y"), ]
pos_JV_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$JV_Flag == "Y"), ]
if(nrow(pre_JV_Pairs) > 0 && nrow(pos_JV_Pairs) > 0){
impact_JV <- Impact_MnA_on_Firms(pre_JV_Pairs, pos_JV_Pairs, "JV_")
impact_general <- dplyr::left_join(impact_general, impact_JV, by = "cusipAup")
}
# dim(impact_general)
# # 2.b RnD network #
pre_RnD_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$RnD_Flag == "Y"), ]
pos_RnD_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$RnD_Flag == "Y"), ]
if(nrow(pre_RnD_Pairs) > 0 && nrow(pos_RnD_Pairs) > 0){
impact_RnD <- Impact_MnA_on_Firms(pre_RnD_Pairs, pos_RnD_Pairs, "RnD_")
impact_general <- dplyr::left_join(impact_general, impact_RnD, by = "cusipAup")
}
# # 2.c Technology network #
pre_Technology_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$Technology_Transfer == "Y"), ]
pos_Technology_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$Technology_Transfer == "Y"), ]
if(nrow(pre_Technology_Pairs) > 0 && nrow(pos_Technology_Pairs) > 0){
impact_Technology <- Impact_MnA_on_Firms(pre_Technology_Pairs, pos_Technology_Pairs, "Technology_")
impact_general <- dplyr::left_join(impact_general, impact_Technology, by = "cusipAup")
}
##
if(nrow(impact_general) > 0){impact_general$event_number <- df$event_number}
print(row_)
print(dim(impact_general))
print(df)
return(impact_general)}
# 2.2
add_centrality_change_2.2 <- function(row_){
# row_ <- 1 # test
df <- MnA[row_, ]
acquir_UP <- df$acquirer.cusip_UP
target_UP <- df$target.cusip_UP
pre_event.date <- as.Date(df$date_ann - 1)
event.date <- as.Date(df$date_ann)
# for test only
# pair_year_updated[which(pair_year_updated$Ali_Ann_Date == event.date),]
# pair_year_updated[which(pair_year_updated$Ali_Expiration_Date == pre_event.date),]
# pair_test <- rbind(pair_year_updated[which(pair_year_updated$Ali_Ann_Date == event.date),], pair_year_updated[which(pair_year_updated$Ali_Expiration_Date == pre_event.date),])
# pair_test[which(pair_test$Ali_Ann_Date <= pre_event.date & pair_test$Ali_Expiration_Date >= pre_event.date), ]
# pair_test[which(pair_test$Ali_Ann_Date <= event.date & pair_test$Ali_Expiration_Date >= event.date ), ]
Impact_MnA_on_Firms <- function(a_pre_network, a_post_network, prefix){
centrality <- function(test){
# test <- pre_Ali_Pairs # test <- pos_Ali_Pairs
a <- test[, c("X1_UP", "X2_UP")]
vec <- c(t(a))
g <- igraph::make_graph(vec, directed = F) %>% igraph::simplify()
adjacency_matrix <- as.matrix(igraph::as_adjacency_matrix(g))
Acquirer_from_Outside <- sum(names(igraph::V(g)) == acquir_UP) == 0
Target_from_Outside <- sum(names(igraph::V(g)) == target_UP) == 0
acquir_position <- if(Acquirer_from_Outside){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[acquir_UP,]}
target_position <- if(Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
adjacency_matrix[target_UP,]}
result <- data.frame(cusipAup = names(igraph::V(g)),
# Add notes
Acquirer_from_Outside = ifelse(Acquirer_from_Outside == TRUE, "Y", "N"),
Target_from_Outside = ifelse(Target_from_Outside == TRUE, "Y", "N"),
# ego network cohesion
ego_nw_cohesion = unlist(purrr::map(igraph::make_ego_graph(g, order = 1, nodes = igraph::V(g)), igraph::cohesion)),
# centrality
degree=igraph::degree(g),
eigen=igraph::eigen_centrality(g)$vector,
btwness=igraph::betweenness(g),
constraint=igraph::constraint(g),
# shortest path (geodesic)
shortest.path.to.acquir = if(Acquirer_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
as.vector(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == acquir_UP)))},
shortest.path.to.target = if(Target_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
as.vector(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == target_UP)))},
shortest.path_acquir_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(Inf, length(names(igraph::V(g))))}else{
rep(igraph::shortest.paths(g, v = which(names(igraph::V(g)) == target_UP))[which(names(igraph::V(g)) == acquir_UP)], length(names(igraph::V(g))))},
# positional similarity (correlation)
positional_correlation_to_acquir = if(Acquirer_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[acquir_UP,]},
positional_correlation_to_target = if(Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[target_UP,]},
positional_correlation_Acquirer_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(0, length(names(igraph::V(g))))}else{
cor(adjacency_matrix)[target_UP,][acquir_UP]},
# positional similarity (euclidean)
positional_euclidean_to_acquir = if(Acquirer_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - acquir_position)^2)))},
positional_euclidean_to_target = if(Target_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - target_position)^2)))},
positional_euclidean_Acquirer_to_target = if(Acquirer_from_Outside | Target_from_Outside){rep(NA, length(names(igraph::V(g))))}else{
apply(adjacency_matrix, 1, function(x) sqrt(sum((x - target_position)^2)))[acquir_UP]},
stringsAsFactors = FALSE)
return(result)
}
# test
# a_pre_network <- pre_Ali_Pairs
# a_post_network <- pos_Ali_Pairs
# prefix <- "GENERAL_"
pre_centrality <- centrality(a_pre_network)
pos_centrality <- centrality(a_post_network)
# replace infinite with max value
pre_centrality$shortest.path.to.acquir[is.infinite(pre_centrality$shortest.path.to.acquir)] <- NA
pos_centrality$shortest.path.to.acquir[is.infinite(pos_centrality$shortest.path.to.acquir)] <- NA
pre_centrality$shortest.path.to.target[is.infinite(pre_centrality$shortest.path.to.target)] <- NA
pos_centrality$shortest.path.to.target[is.infinite(pos_centrality$shortest.path.to.target)] <- NA
pre_centrality$shortest.path_acquir_to_target[is.infinite(pre_centrality$shortest.path_acquir_to_target)] <- NA
pos_centrality$shortest.path_acquir_to_target[is.infinite(pos_centrality$shortest.path_acquir_to_target)] <- NA
max.value <- max(max(pre_centrality$shortest.path.to.acquir, na.rm = TRUE),
max(pre_centrality$shortest.path.to.target, na.rm = TRUE),
max(pos_centrality$shortest.path.to.acquir, na.rm = TRUE),
max(pos_centrality$shortest.path.to.target, na.rm = TRUE))
pre_centrality$shortest.path.to.acquir[is.na(pre_centrality$shortest.path.to.acquir)] <- max.value + 1
pos_centrality$shortest.path.to.acquir[is.na(pos_centrality$shortest.path.to.acquir)] <- max.value + 1
pre_centrality$shortest.path.to.target[is.na(pre_centrality$shortest.path.to.target)] <- max.value + 1
pos_centrality$shortest.path.to.target[is.na(pos_centrality$shortest.path.to.target)] <- max.value + 1
pre_centrality$shortest.path_acquir_to_target[is.na(pre_centrality$shortest.path_acquir_to_target)] <- max.value + 1
pos_centrality$shortest.path_acquir_to_target[is.na(pos_centrality$shortest.path_acquir_to_target)] <- max.value + 1
max.value_euc <- max(max(pre_centrality$positional_euclidean_to_acquir, na.rm = TRUE),
max(pre_centrality$positional_euclidean_to_target, na.rm = TRUE),
max(pos_centrality$positional_euclidean_to_acquir, na.rm = TRUE),
max(pos_centrality$positional_euclidean_to_target, na.rm = TRUE))
pre_centrality$positional_euclidean_to_acquir[is.na(pre_centrality$positional_euclidean_to_acquir)] <- max.value_euc + 1
pre_centrality$positional_euclidean_to_target[is.na(pre_centrality$positional_euclidean_to_target)] <- max.value_euc + 1
pre_centrality$positional_euclidean_Acquirer_to_target[is.na(pre_centrality$positional_euclidean_Acquirer_to_target)] <- max.value_euc + 1
pos_centrality$positional_euclidean_to_acquir[is.na(pos_centrality$positional_euclidean_to_acquir)] <- max.value_euc + 1
pos_centrality$positional_euclidean_to_target[is.na(pos_centrality$positional_euclidean_to_target)] <- max.value_euc + 1
pos_centrality$positional_euclidean_Acquirer_to_target[is.na(pos_centrality$positional_euclidean_Acquirer_to_target)] <- max.value_euc + 1
names(pre_centrality)[2:17] <- paste0("pre_", names(pre_centrality)[2:17])
# delta centrality #
merged_ <- merge(pre_centrality, pos_centrality, by = "cusipAup")
merged_$pre_btwness <- regrrr::scale_01(merged_$pre_btwness)
merged_$btwness <- regrrr::scale_01(merged_$btwness)
merged_$delta_degree <- merged_$degree - merged_$pre_degree
merged_$delta_eigen <- merged_$eigen - merged_$pre_eigen
merged_$delta_btwness <- merged_$btwness - merged_$pre_btwness
merged_$delta_strhole <- merged_$pre_constraint - merged_$constraint
merged_$delta_nw_cohesion <- merged_$ego_nw_cohesion - merged_$pre_ego_nw_cohesion
merged_$delta_spath2acquir <- merged_$shortest.path.to.acquir - merged_$pre_shortest.path.to.acquir
merged_$delta_spath2target <- merged_$shortest.path.to.target - merged_$pre_shortest.path.to.target
# delta position #
merged_$delta_cor_to_acquir <- merged_$positional_correlation_to_acquir - merged_$pre_positional_correlation_to_acquir
merged_$delta_cor_to_target <- merged_$positional_correlation_to_target - merged_$pre_positional_correlation_to_target
merged_$delta_euc_to_acquir <- merged_$positional_euclidean_to_acquir - merged_$pre_positional_euclidean_to_acquir
merged_$delta_euc_to_target <- merged_$positional_euclidean_to_target - merged_$pre_positional_euclidean_to_target
merged_ <- merged_ %>% dplyr::filter(abs(delta_eigen) > 0.001 | abs(delta_strhole) > 0.001)
names(merged_)[2:length(names(merged_))] <- paste0(prefix, names(merged_)[2:length(names(merged_))])
return(merged_)}
# 1. general alliance network #
pre_Ali_Pairs <- pair_year_updated[which(pair_year_updated$Ali_Ann_Date <= pre_event.date & pair_year_updated$Ali_Expiration_Date >= pre_event.date), ]
pos_Ali_Pairs <- pair_year_updated[which(pair_year_updated$Ali_Ann_Date <= event.date & pair_year_updated$Ali_Expiration_Date >= event.date ), ]
if(nrow(pre_Ali_Pairs) > 0 && nrow(pos_Ali_Pairs) > 0){
impact_general <- Impact_MnA_on_Firms(pre_Ali_Pairs, pos_Ali_Pairs, "GENERAL_")
}
# dim(impact_general)
# # 2.a JV network #
pre_JV_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$JV_Flag == "Y"), ]
pos_JV_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$JV_Flag == "Y"), ]
if(nrow(pre_JV_Pairs) > 0 && nrow(pos_JV_Pairs) > 0){
impact_JV <- Impact_MnA_on_Firms(pre_JV_Pairs, pos_JV_Pairs, "JV_")
impact_general <- dplyr::left_join(impact_general, impact_JV, by = "cusipAup")
}
# dim(impact_general)
# # 2.b RnD network #
pre_RnD_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$RnD_Flag == "Y"), ]
pos_RnD_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$RnD_Flag == "Y"), ]
if(nrow(pre_RnD_Pairs) > 0 && nrow(pos_RnD_Pairs) > 0){
impact_RnD <- Impact_MnA_on_Firms(pre_RnD_Pairs, pos_RnD_Pairs, "RnD_")
impact_general <- dplyr::left_join(impact_general, impact_RnD, by = "cusipAup")
}
# # 2.c Technology network #
pre_Technology_Pairs <- pre_Ali_Pairs[which(pre_Ali_Pairs$Technology_Transfer == "Y"), ]
pos_Technology_Pairs <- pos_Ali_Pairs[which(pos_Ali_Pairs$Technology_Transfer == "Y"), ]
if(nrow(pre_Technology_Pairs) > 0 && nrow(pos_Technology_Pairs) > 0){
impact_Technology <- Impact_MnA_on_Firms(pre_Technology_Pairs, pos_Technology_Pairs, "Technology_")
impact_general <- dplyr::left_join(impact_general, impact_Technology, by = "cusipAup")
}
##
if(nrow(pre_Ali_Pairs) > 0){
if(nrow(impact_general) > 0){impact_general$impacting_or_not <- df$impacting_or_not}
}
print(row_)
return(impact_general)}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.